--- schema_version: 1 rubric_version: "1.1" kind: pair canonical_url: "https://onaihq.com/compare/agents/swissgpt-vs-typingmind/" data_url: "https://onaihq.com/compare/agents/swissgpt-vs-typingmind/data.json" generated_at: 2026-10-27 last_checked_at: 2026-10-26 last_cross_checked_at: 2026-10-26 next_check_at: 2026-12-20 publisher: "onAI" --- # AlpineAI / SwissGPT vs TypingMind / TypingMind Teams > Published by onAI. onAI also sells one of the products compared here. Findings about other products link to public sources so you can check them; statements about onAI are marked as stated by onAI. This comparison has 28 criteria with sourced findings for both products, including 28 that a second agent cross-checked for both. 51 criteria have a sourced finding for at least one product. "Sourced" findings passed publication validation: they are scoped, fresh and cite public resolved sources. "Agent cross-checked" findings also received a separate agent source pass, not necessarily a hands-on test. "Stated by onAI" marks the publisher's own statements about its product (evidence type vendor-statement), which cite no third-party source. Humans approve the rubric, templates and generator; they do not review every finding. It does not declare a winner. A missing sourced finding is not evidence that a capability is absent. - AlpineAI / SwissGPT source checks: 2026-10-21 to 2026-10-26. - TypingMind / TypingMind Teams source checks: 2026-10-21 to 2026-10-26. - Canonical HTML: https://onaihq.com/compare/agents/swissgpt-vs-typingmind/ - Machine-readable JSON: https://onaihq.com/compare/agents/swissgpt-vs-typingmind/data.json - Methodology: https://onaihq.com/compare/agents/methodology/ ## Models ### MOD-01: Model catalogue and providers Judgment rule: List each available model by provider and plan, including reasoning, fast and multimodal variants. Record whether the user picks the model, the product routes automatically, or both. A provider counts only if at least one of its current models is selectable on the scoped plan. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial AlpineAI names Mistral Large 3 as its primary model and says it operates all models itself in Switzerland. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - The FAQ does not enumerate all models, expose a per-chat picker, explain routing visibility or map multimodal inputs to models. - providers — Which model providers are available? [partial]: Mistral is the named current provider. Other providers are not enumerated. Sources: swissgpt:swissgpt:product. - current_models — Which current models are selectable on this plan? [partial]: Mistral Large 3 is named as the primary model; this does not establish a user-selectable catalogue. Sources: swissgpt:swissgpt:product. - user_choice — Can users pick the model per chat? [unknown]: Can users pick the model per chat? Unknown. The FAQ does not enumerate all models, expose a per-chat picker, explain routing visibility or map multimodal inputs to models. - auto_routing — Does the product route between models automatically, and is the chosen model shown? [unknown]: Does the product route between models automatically, and is the chosen model shown? Unknown. The FAQ does not enumerate all models, expose a per-chat picker, explain routing visibility or map multimodal inputs to models. - multimodal_input — Which models accept images, audio or files? [unknown]: Which models accept images, audio or files? Unknown. The FAQ does not enumerate all models, expose a per-chat picker, explain routing visibility or map multimodal inputs to models. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:product #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Teams is a customer-key multi-provider chat interface, with dated additions and explicit model selection. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=TypingMind Teams SaaS; Starter, Growth and Professional prerequisites stated per answer; Business quote not obtained; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - providers — Which model providers are available? [supported]: Teams admin key setup lists OpenAI, Anthropic, Google, OpenRouter, DeepInfra, DeepSeek, Groq, Mistral, Moonshot, Perplexity, xAI and Z.ai, plus custom providers. Sources: typingmind:typingmind:set-up-api-keys. - current_models — Which current models are selectable on this plan? [partial]: The dated October changelog records Gemini 4 Argon and Muse Wonder on October 14 and Claude Haiku 5.5, Mistral Large 4 and Nano Banana 2.1 on October 9. Actual availability depends on enabled endpoint and customer account; a complete Teams tenant model catalogue was not observed. Sources: typingmind:typingmind:release, typingmind:typingmind:set-up-api-keys. - user_choice — Can users pick the model per chat? [supported]: Users select an available model for a chat; Teams administrators choose which models are exposed. Sources: typingmind:typingmind:manage-ai-models, typingmind:typingmind:restrict-model-access. - auto_routing — Does the product route between models automatically, and is the chosen model shown? [unknown]: Does the product route between models automatically, and is the chosen model shown? The first-party Teams documentation, pricing, legal pages and dated changelog retrieved on 2026-10-21 do not establish this detail; no account or bounded fixture was used. - multimodal_input — Which models accept images, audio or files? [partial]: Document upload, vision-capable model image input and transcription are documented as separate routes; the docs do not provide a complete per-model Teams image/audio/file matrix. Sources: typingmind:typingmind:chat-with-documents, typingmind:typingmind:voice. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:set-up-api-keys, typingmind:typingmind:release, typingmind:typingmind:manage-ai-models, typingmind:typingmind:restrict-model-access, typingmind:typingmind:chat-with-documents, typingmind:typingmind:voice ### MOD-02: Model freshness Judgment rule: Measure the days from a provider's general release to availability in the product for the last three major releases from OpenAI, Anthropic and Google (and Mistral where offered). Use dated changelogs or observation. Record deprecation notice for retired models. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Dated model additions are public; release lag and removal guarantees remain unestablished. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=TypingMind Teams SaaS; Starter, Growth and Professional prerequisites stated per answer; Business quote not obtained; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - release_lag_days — What was the lag for each of the last three major releases, per provider? [unknown]: What was the lag for each of the last three major releases, per provider? No matched first-party provider release dates for the last three releases, fixed deprecation-notice period or version-retention guarantee was retrieved. - changelog — Is there a dated public model changelog? [supported]: The public changelog dates model and Teams feature releases through October 14, 2026 in the retrieved page. Sources: typingmind:typingmind:release. - deprecation_notice — How much notice is given before a model is removed? [unknown]: How much notice is given before a model is removed? No matched first-party provider release dates for the last three releases, fixed deprecation-notice period or version-retention guarantee was retrieved. - version_pinning — Can an admin keep an older model version? [partial]: Customer custom-model endpoint configuration can name a provider model or deployment, but a contractual right to retain discontinued SaaS wrappers or provider versions is not established. Sources: typingmind:typingmind:manage-ai-models, typingmind:typingmind:azure-openai. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:release, typingmind:typingmind:manage-ai-models, typingmind:typingmind:azure-openai ### MOD-03: Bring-your-own keys and provider contracts Judgment rule: Record customer-held API keys, customer cloud accounts (Azure OpenAI, Bedrock, Vertex), and aggregator routes (OpenRouter, LiteLLM) as separate routes. Record what still works without a vendor-provided model, whether prices change with BYO, and documented restrictions. API-key support does not imply support for the customer's own Azure contract. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Teams uses customer provider keys; its Custom Admin documentation also describes Claude via the customer AWS Bedrock account. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel, Commercial self-hosted TypingMind Teams server; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams SaaS Starter/Growth/Professional custom-model setup; Bedrock Custom Admin route has no verified exact tier mapping. Separate Teams self-host licence for private endpoint configuration.; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=not specified; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Bedrock is documented for Custom Admin; no all-tier, all-Claude-model or self-host Bedrock entitlement is inferred. Provider-route feature parity was not tested. - own_api_keys — Can the customer use its own provider API keys? [supported]: Admins preset the organization’s provider keys. End users cannot supply separate personal keys in the Teams chat UI. Sources: typingmind:typingmind:set-up-api-keys. - own_cloud_contract — Can it use its own Azure OpenAI, Bedrock or Vertex deployment? [partial]: Teams documents Azure OpenAI custom-model setup. Its AWS Bedrock guide explicitly applies to TypingMind Custom Admin at custom.typingmind.com/manage and configures Claude via Bedrock using the customer AWS keys, region and approved model access. Exact Bedrock tier/current-model coverage and Vertex support remain unknown. Sources: typingmind:typingmind:r2-teams-custom-models, typingmind:typingmind:r3-custom-admin-bedrock, typingmind:typingmind:r2-plan-matrix. - custom_endpoint — Can an admin add any OpenAI-compatible endpoint? [partial]: Teams Admin Panel supports Add Custom Model, naming Azure OpenAI, Hugging Face and Ollama. This does not establish an arbitrary-endpoint compatibility guarantee. Separately, commercial Teams self-host operators can enable otherwise-blocked private/localhost custom endpoints; connectivity and model feature parity were not tested. Sources: typingmind:typingmind:r2-teams-custom-models, typingmind:typingmind:r2-plan-matrix, typingmind:typingmind:get-started-with-self-host-installation-instruct. - byo_pricing — Does BYO change the licence price? [supported]: Teams subscription excludes API usage and requires customer keys; BYO is the standard paid model route, not an evidenced licence discount. Sources: typingmind:typingmind:pricing, typingmind:typingmind:set-up-api-keys. - byo_restrictions — Which features stop working with BYO? [partial]: Capabilities depend on the selected model and provider API supporting files, tool use and related features; there is no complete feature-by-provider parity matrix. Sources: typingmind:typingmind:manage-ai-models, typingmind:typingmind:toolbox. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:set-up-api-keys, typingmind:typingmind:r2-teams-custom-models, typingmind:typingmind:r3-custom-admin-bedrock, typingmind:typingmind:r2-plan-matrix, typingmind:typingmind:get-started-with-self-host-installation-instruct, typingmind:typingmind:pricing, typingmind:typingmind:manage-ai-models, typingmind:typingmind:toolbox ### MOD-04: Local and self-hosted models Judgment rule: Record support for models running in the customer's environment (vLLM, Ollama, TGI or similar) and which features work with them: chat, RAG, tools, embeddings. Distinguish a documented, supported route from a generic endpoint that happens to work. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial AlpineAI says it hosts open models on its own Swiss hardware. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - Customer-hosted chat and embedding routes, supported runtimes and their feature parity are not documented in the reviewed current offer. - self_hosted_chat — Can chat use a customer-hosted model? [unknown]: Can chat use a customer-hosted model? Unknown. Customer-hosted chat and embedding routes, supported runtimes and their feature parity are not documented in the reviewed current offer. - self_hosted_embeddings — Can retrieval use customer-hosted embeddings? [unknown]: Can retrieval use customer-hosted embeddings? Unknown. Customer-hosted chat and embedding routes, supported runtimes and their feature parity are not documented in the reviewed current offer. - feature_parity — Which features fail with self-hosted models? [unknown]: Which features fail with self-hosted models? Unknown. Customer-hosted chat and embedding routes, supported runtimes and their feature parity are not documented in the reviewed current offer. - vendor_hosted_open_models — Does the vendor host open-weight models in EU/CH? [supported]: The vendor identifies Mistral Large 3 as its current primary open model, operated on AlpineAI-owned hardware in Switzerland without foreign model APIs. Sources: swissgpt:swissgpt:product. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:product #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Customer model endpoints are supported; optional self-host knowledge uses documented OpenAI embeddings. Availability: generally-available Delivery: native Scope: surfaces=TypingMind personal browser app and PWA, TypingMind Teams SaaS chat and Admin Panel, Commercial self-hosted TypingMind Teams server; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Personal app local-model guides; Teams SaaS custom-model setup on Starter/Growth/Professional; separate annual commercial Teams self-host package; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=not specified; region=Personal browser/local network; US/EU Teams SaaS; customer-selected self-host placement and external API routes; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - self_hosted_chat — Can chat use a customer-hosted model? [partial]: Personal typingmind.com has Ollama and LM Studio guides. Teams-specific Admin Panel instructions independently name Ollama custom models, but do not establish the personal local-connection procedure for Teams SaaS. Commercial Teams self-host can explicitly enable private endpoints; no hosted-to-local connectivity fixture was run. Sources: typingmind:typingmind:r2-teams-custom-models, typingmind:typingmind:r2-plan-matrix, typingmind:typingmind:ollama, typingmind:typingmind:lmstudio, typingmind:typingmind:get-started-with-self-host-installation-instruct. - self_hosted_embeddings — Can retrieval use customer-hosted embeddings? [partial]: The optional KB module in the separately quoted annual Teams self-host package requires an OpenAI embeddings key and PostgreSQL pgvector. A customer-hosted embedding substitution is unverified. This is separate from personal local chat models. Sources: typingmind:typingmind:get-started-with-self-host-installation-instruct, typingmind:typingmind:cloud-host-vs-self-host. - feature_parity — Which features fail with self-hosted models? [partial]: Toolbox needs a model capable of calling tools; other multimodal capabilities depend on the custom model. Complete parity and failure behaviour are undocumented. Sources: typingmind:typingmind:toolbox, typingmind:typingmind:manage-ai-models. - vendor_hosted_open_models — Does the vendor host open-weight models in EU/CH? [unknown]: Does the vendor host open-weight models in EU/CH? No vendor-hosted open-weight EU/CH model service, supported embedding substitution or full self-host-model parity matrix was established. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:r2-teams-custom-models, typingmind:typingmind:r2-plan-matrix, typingmind:typingmind:ollama, typingmind:typingmind:lmstudio, typingmind:typingmind:get-started-with-self-host-installation-instruct, typingmind:typingmind:cloud-host-vs-self-host, typingmind:typingmind:toolbox, typingmind:typingmind:manage-ai-models ### MOD-05: Model control and usage limits Judgment rule: Record admin allow-lists per group, default models, restrictions on which models may see which data, and per-user rate or fair-use limits. Test one restriction with two users in different groups. Record limits as enforced stops, warnings or undocumented fair use. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Professional administrators govern model access and configurable consumption limits. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams Professional for governance; customer-configured limits; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - model_allowlist_by_group — Can admins limit models per group? [supported]: Admins restrict model availability by user group; the limits guide still points availability controls to the legacy model-limits section. Sources: typingmind:typingmind:restrict-model-access, typingmind:typingmind:usage-limits. - data_bound_models — Can admins restrict which models receive connector or knowledge content? [partial]: Admins configure the models available to agents and their knowledge access. No content-classification or connector-label-based model egress rule was established. Sources: typingmind:typingmind:roles-and-permissions, typingmind:typingmind:connect-knowledge-base-to-your-ai-agents. - default_model — Can admins set the default model? [unknown]: No Teams administrator control setting the default model for new chats was established. The October multi-model finalization default is a different feature; the plan matrix and model-management documents do not resolve this atom. - user_limits — What per-user message or token limits apply, and are they published? [supported]: Reusable limit groups configure message, character and token limits per time period, across users/groups, models, agents and plugins. Published examples are admin settings, not plan-wide quotas. Sources: typingmind:typingmind:usage-limits. - limit_behaviour — What happens when a limit is reached? [partial]: Users see a configurable limitation notice and can inspect details. Exact block, downgrade, overshoot and reset semantics at exhaustion were not established. Sources: typingmind:typingmind:usage-limits. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:restrict-model-access, typingmind:typingmind:usage-limits, typingmind:typingmind:roles-and-permissions, typingmind:typingmind:connect-knowledge-base-to-your-ai-agents ## Company knowledge ### KNOW-01: Files in chat Judgment rule: Record supported formats, size and count limits, scanned-PDF OCR, tables and spreadsheets, images inside documents, and whether a file is read in full or retrieved in chunks. Upload the shared fixture set (long German PDF, scanned contract, Excel with formulas, PowerPoint) and record failures. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial Document upload is advertised for assistants and chat. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - The published 50-document/100 MB assistant limit does not define per-file and per-chat limits. OCR, embedded images, formula/table fidelity and full-file versus chunked processing need technical documentation or the shared upload fixture. - formats — Which formats are accepted? [partial]: PDF, Word, text and Excel are named across the product and sector pages. This is not an exhaustive format matrix. Sources: swissgpt:swissgpt:product, swissgpt:swissgpt:university. - limits — What size and count limits apply per file and chat? [unknown]: What size and count limits apply per file and chat? Unknown. The published 50-document/100 MB assistant limit does not define per-file and per-chat limits. OCR, embedded images, formula/table fidelity and full-file versus chunked processing need technical documentation or the shared upload fixture. - ocr — Are scanned PDFs read? [unknown]: Are scanned PDFs read? Unknown. The published 50-document/100 MB assistant limit does not define per-file and per-chat limits. OCR, embedded images, formula/table fidelity and full-file versus chunked processing need technical documentation or the shared upload fixture. - spreadsheets — Are Excel tables read with structure intact? [unknown]: Are Excel tables read with structure intact? Unknown. The published 50-document/100 MB assistant limit does not define per-file and per-chat limits. OCR, embedded images, formula/table fidelity and full-file versus chunked processing need technical documentation or the shared upload fixture. - full_vs_chunked — Is the whole file used or only retrieved parts, and does the user see which? [unknown]: Is the whole file used or only retrieved parts, and does the user see which? Unknown. The published 50-document/100 MB assistant limit does not define per-file and per-chat limits. OCR, embedded images, formula/table fidelity and full-file versus chunked processing need technical documentation or the shared upload fixture. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:product, swissgpt:swissgpt:university #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Document text can be attached to chat; full-context training files and RAG are distinct. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=TypingMind Teams SaaS; Starter, Growth and Professional prerequisites stated per answer; Business quote not obtained; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - formats — Which formats are accepted? [supported]: The chat guide lists PDF, Markdown, CSV, XLSX, TXT, DOCX, JSON and HTML. Images require a suitable vision model. Sources: typingmind:typingmind:chat-with-documents. - limits — What size and count limits apply per file and chat? [partial]: Chat uploads are bounded by the selected model’s context, but a numeric per-chat file-count/size limit was not established. The separate KB documents 50MB uploads. Sources: typingmind:typingmind:chat-with-documents, typingmind:typingmind:how-knowledge-base-works-in-typingmind-team. - ocr — Are scanned PDFs read? [unknown]: Are scanned PDFs read? No scanned-PDF OCR accuracy, spreadsheet-preservation fixture or numerical ordinary-chat upload matrix was established. - spreadsheets — Are Excel tables read with structure intact? [partial]: XLSX and CSV are accepted; Toolbox executes local Python/JavaScript analysis. Preservation of merged cells, formula results and workbook structure was not tested. Sources: typingmind:typingmind:chat-with-documents, typingmind:typingmind:toolbox. - full_vs_chunked — Is the whole file used or only retrieved parts, and does the user see which? [supported]: Ordinary chat document text enters chat context and can later leave the model window; agent training files are injected into system context, while KB access retrieves relevant chunks. These are separate modes. Sources: typingmind:typingmind:chat-with-documents, typingmind:typingmind:training-files-vs-knowledge-base. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:chat-with-documents, typingmind:typingmind:how-knowledge-base-works-in-typingmind-team, typingmind:typingmind:toolbox, typingmind:typingmind:training-files-vs-knowledge-base ### KNOW-02: Knowledge bases and shared collections Judgment rule: Record curated document collections that persist across chats, who can create and share them, size limits, and how they attach to assistants. Distinguish personal, group and organisation-wide scope. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial Standard and Premium include document-backed assistants with published collection limits. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - User/group/organisation sharing and Enterprise collection limits are not specified. The wording does not clearly distinguish per-file from aggregate 100 MB. - persistent_collections — Can users keep document collections across chats? [supported]: Assistants retain their own reference-document workspace for later questions. Sources: swissgpt:swissgpt:product. - sharing_scope — Can collections be shared with users, groups or the whole organisation? [unknown]: Can collections be shared with users, groups or the whole organisation? Unknown. User/group/organisation sharing and Enterprise collection limits are not specified. The wording does not clearly distinguish per-file from aggregate 100 MB. - size_limits — What are the collection limits? [partial]: Standard has 2 assistants and Premium 5; each is listed as 50 documents / max. 100 MB. Enterprise limits are not published. Sources: swissgpt:swissgpt:university. - assistant_binding — Can collections be bound to an assistant? [supported]: Users place documents in an assistant workspace and questions use those sources. Sources: swissgpt:swissgpt:product. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:product, swissgpt:swissgpt:university #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Growth and above provide a tagged organization KB bound to agents; the technical guide still labels it Early Access. Availability: beta Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Growth/Professional: RAG and tagged agent KB; Professional additionally supplies group/resource visibility restrictions; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Current plan matrix sells RAG in Growth/Professional, while the technical KB guide (updated2025-11-10) still labels it Early Access. No dated GA transition or tenant rollout was verified; findings retain beta/Early Access qualification. - No source-user permission fixture was run. - persistent_collections — Can users keep document collections across chats? [supported]: The organization KB persists documents and connected data across conversations; this differs from temporary chat uploads and full-context agent training files. Sources: typingmind:typingmind:get-started-with-knowledge-base, typingmind:typingmind:training-files-vs-knowledge-base. - sharing_scope — Can collections be shared with users, groups or the whole organisation? [partial]: Growth/Professional support tagged KB data bound to agents. Per-user/group agent restrictions require Professional governance; these restrictions do not establish per-document source ACL mirroring. Sources: typingmind:typingmind:connect-knowledge-base-to-your-ai-agents, typingmind:typingmind:restrict-ai-agents-access, typingmind:typingmind:r2-plan-matrix. - size_limits — What are the collection limits? [supported]: The KB guide states 50MB per upload and one million included training characters with extra capacity purchasable; indexed chunks are about 3,000 words and retrieval is up to five chunks. Sources: typingmind:typingmind:how-knowledge-base-works-in-typingmind-team. - assistant_binding — Can collections be bound to an assistant? [supported]: Knowledge Base Access in an agent selects tags identifying the accessible data. Sources: typingmind:typingmind:connect-knowledge-base-to-your-ai-agents. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:get-started-with-knowledge-base, typingmind:typingmind:training-files-vs-knowledge-base, typingmind:typingmind:connect-knowledge-base-to-your-ai-agents, typingmind:typingmind:restrict-ai-agents-access, typingmind:typingmind:r2-plan-matrix, typingmind:typingmind:how-knowledge-base-works-in-typingmind-team ### KNOW-03: Enterprise connectors Judgment rule: List each connector separately with plan, availability and setup owner: SharePoint, OneDrive, Outlook, Teams, Google Drive, Gmail, Confluence, Jira, Notion, Box, Salesforce, ServiceNow, file shares. Record whether the vendor or a third party (or the customer via MCP) provides it, and whether it needs admin consent. A connector listed on a roadmap or partner page is `announced`. An on-prem connector or gateway proves access to local data, not a self-hosted platform (DATA-02). #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial Enterprise sector pages name production REST integrations with eGeKo and five healthcare systems. A public standard connector matrix for M365, Google, Atlassian and on-premise sources remains unverified. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - The production integration names are vendor claims on the sector pages, not a tested connector inventory. - Individual integrations require the Enterprise package from 10 licences; no API account, customer deployment or fixture was accessed. - microsoft_365 — Which M365 sources (SharePoint, OneDrive, Outlook, Teams) are connected? [unknown]: Which M365 sources (SharePoint, OneDrive, Outlook, Teams) are connected? Unknown. The current FAQ does not identify shipped SharePoint, OneDrive, Outlook-search, Teams, Google, Atlassian or on-prem-source connectors, their consent scopes or maintainers. The Outlook add-in is recorded under WORK-02. Rechecked 2026-10-26 in AlpineAI product, public-administration and healthcare feature/licence FAQs and Security page. No extra technical/admin documentation was identified in official links or targeted first-party searches; no account fixture was run. - google_workspace — Which Google sources are connected? [unknown]: Which Google sources are connected? Unknown. The current FAQ does not identify shipped SharePoint, OneDrive, Outlook-search, Teams, Google, Atlassian or on-prem-source connectors, their consent scopes or maintainers. The Outlook add-in is recorded under WORK-02. Rechecked 2026-10-26 in AlpineAI product, public-administration and healthcare feature/licence FAQs and Security page. No extra technical/admin documentation was identified in official links or targeted first-party searches; no account fixture was run. - atlassian — Confluence and Jira? [unknown]: Confluence and Jira? Unknown. The current FAQ does not identify shipped SharePoint, OneDrive, Outlook-search, Teams, Google, Atlassian or on-prem-source connectors, their consent scopes or maintainers. The Outlook add-in is recorded under WORK-02. Rechecked 2026-10-26 in AlpineAI product, public-administration and healthcare feature/licence FAQs and Security page. No extra technical/admin documentation was identified in official links or targeted first-party searches; no account fixture was run. - other_systems — Which other business systems? [partial]: The public-administration FAQ names eGeKo as an integration in production. The healthcare FAQ names Axenita, Aeskulap, Ines, Logival and Mediway as production integrations through the AlpineAI REST API. Individual integrations are part of Enterprise from 10 licences and planned with the customer IT team. Public consent scopes, per-system configuration and a self-service connector matrix are not supplied. Sources: swissgpt:swissgpt:public-pricing-20261026, swissgpt:swissgpt:health-pricing-20261026. - on_prem_sources — Can on-prem file shares or databases be connected? [unknown]: Can on-prem file shares or databases be connected? Unknown. The current FAQ does not identify shipped SharePoint, OneDrive, Outlook-search, Teams, Google, Atlassian or on-prem-source connectors, their consent scopes or maintainers. The Outlook add-in is recorded under WORK-02. Rechecked 2026-10-26 in AlpineAI product, public-administration and healthcare feature/licence FAQs and Security page. No extra technical/admin documentation was identified in official links or targeted first-party searches; no account fixture was run. - connector_provider — Who builds and maintains each connector? [partial]: AlpineAI plans custom Enterprise integrations with the institution IT team. The university FAQ mentions expanding prebuilt connectors without naming them. Sources: swissgpt:swissgpt:university. Checked 2026-10-26; agent cross-checked 2026-10-26; check again by 2026-12-25. Sources: swissgpt:swissgpt:university, swissgpt:swissgpt:public-pricing-20261026, swissgpt:swissgpt:health-pricing-20261026, swissgpt:swissgpt:product-20261026, swissgpt:swissgpt:security-20261026 #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Named knowledge integrations and extensible tools exist; broad source mentions lack enterprise permission detail. Availability: beta Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel, Commercial self-hosted TypingMind Teams server; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Growth/Professional SaaS RAG; separate annual commercial Teams self-host licence with optional KB and customer-managed connector credentials; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=not specified; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Current plan matrix sells RAG in Growth/Professional, while the technical KB guide (updated2025-11-10) still labels it Early Access. No dated GA transition or tenant rollout was verified; findings retain beta/Early Access qualification. - A reachable self-host plugin endpoint is not a native file-share crawler. - microsoft_365 — Which M365 sources (SharePoint, OneDrive, Outlook, Teams) are connected? [partial]: The training-files comparison mentions OneDrive as a possible KB source; no retrieved setup guide establishes SharePoint, Outlook or Teams ingestion, permissions or maintenance. Sources: typingmind:typingmind:training-files-vs-knowledge-base. - google_workspace — Which Google sources are connected? [partial]: Growth/Professional SaaS RAG names Google Drive. In the separate commercial Teams self-host package, external KB connectors require the operator’s OAuth configuration. Gmail and the full Workspace source set remain unverified. Sources: typingmind:typingmind:get-started-with-knowledge-base, typingmind:typingmind:setup-external-connectors-for-knowledge-base, typingmind:typingmind:r2-plan-matrix. - atlassian — Confluence and Jira? [unknown]: Confluence and Jira? No first-party Confluence/Jira connector setup, complete M365 connector matrix or maintained on-prem file-share connector was established. - other_systems — Which other business systems? [supported]: KB documentation names Notion and GitHub; admins select Notion pages/databases. Separate MCP/HTTP plugins can query other systems without proving a maintained KB connector. Sources: typingmind:typingmind:get-started-with-knowledge-base, typingmind:typingmind:connect-with-your-notion-pages, typingmind:typingmind:mcp. - on_prem_sources — Can on-prem file shares or databases be connected? [partial]: Self-host operators can explicitly permit private custom-model/plugin/agent endpoints; custom database actions are described. A supported file-share crawler is not established. Sources: typingmind:typingmind:get-started-with-self-host-installation-instruct, typingmind:typingmind:overview. - connector_provider — Who builds and maintains each connector? [partial]: TypingMind documents built-in KB setup; external self-host sources need operator-managed OAuth credentials and Firecrawl for web scraping. MCP servers and custom plugins have separate maintainers. Sources: typingmind:typingmind:setup-external-connectors-for-knowledge-base, typingmind:typingmind:mcp. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:training-files-vs-knowledge-base, typingmind:typingmind:get-started-with-knowledge-base, typingmind:typingmind:setup-external-connectors-for-knowledge-base, typingmind:typingmind:r2-plan-matrix, typingmind:typingmind:connect-with-your-notion-pages, typingmind:typingmind:mcp, typingmind:typingmind:get-started-with-self-host-installation-instruct, typingmind:typingmind:overview ### KNOW-04: Permission-aware retrieval Judgment rule: Using two test users, verify that a restricted document in the fixture tenant is never cited or summarised for the user without access, including after a permission change. Record whether permissions are enforced at query time by the source system, mirrored into an index, or not enforced. Record the propagation delay of a revocation. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Documented access control is by tagged knowledge and agent/group visibility; per-user source ACL enforcement is unproven. Availability: beta Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Growth/Professional RAG data/tag selection; Professional adds group-based resource restrictions; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Current plan matrix sells RAG in Growth/Professional, while the technical KB guide (updated2025-11-10) still labels it Early Access. No dated GA transition or tenant rollout was verified; findings retain beta/Early Access qualification. - No two-user source-permission or revocation fixture. - source_permissions_respected — Are source permissions enforced per user? [unknown]: Are source permissions enforced per user? No two-user permission/revocation fixture, source ACL mirror, revocation SLA or Microsoft Purview-label enforcement documentation was retrieved. Do not equate agent access with source-user permissions. - enforcement_point — Does the source system, a mirrored index or nothing enforce them? [partial]: Growth/Professional bind data tags to agents; Professional adds agent availability by user/group. This application-side mechanism does not establish synchronized source-user ACLs. Sources: typingmind:typingmind:connect-knowledge-base-to-your-ai-agents, typingmind:typingmind:restrict-ai-agents-access, typingmind:typingmind:r2-plan-matrix. - revocation_delay — How quickly does a removed permission take effect? [unknown]: How quickly does a removed permission take effect? No two-user permission/revocation fixture, source ACL mirror, revocation SLA or Microsoft Purview-label enforcement documentation was retrieved. Do not equate agent access with source-user permissions. - admin_scope_controls — Can admins exclude sites, folders or labels? [supported]: Admins choose the Notion pages/databases to ingest and select knowledge tags for each agent. Sources: typingmind:typingmind:connect-with-your-notion-pages, typingmind:typingmind:connect-knowledge-base-to-your-ai-agents. - sensitivity_labels — Are Microsoft Purview or similar labels respected? [unknown]: Are Microsoft Purview or similar labels respected? No two-user permission/revocation fixture, source ACL mirror, revocation SLA or Microsoft Purview-label enforcement documentation was retrieved. Do not equate agent access with source-user permissions. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:connect-knowledge-base-to-your-ai-agents, typingmind:typingmind:restrict-ai-agents-access, typingmind:typingmind:r2-plan-matrix, typingmind:typingmind:connect-with-your-notion-pages ### KNOW-05: Freshness and sync model Judgment rule: Record live (federated) search versus a copied index, sync interval, deletion propagation, and where the index is stored. Edit and delete one fixture document and record when answers change. A copied index adds storage location to DATA-01. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial The KB indexes copies, refreshes connected content daily and uses selected SaaS or customer-hosted storage. Availability: beta Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel, Commercial self-hosted TypingMind Teams server; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Growth/Professional SaaS RAG; separately quoted annual Teams self-host licence and optional PostgreSQL KB; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=not specified; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Current plan matrix sells RAG in Growth/Professional, while the technical KB guide (updated2025-11-10) still labels it Early Access. No dated GA transition or tenant rollout was verified; findings retain beta/Early Access qualification. - No synchronization/deletion latency fixture. - live_or_indexed — Is content searched live or copied into an index? [supported]: RAG splits and embeds copied source text in a knowledge index; full-context training files and live custom tools are separate. Sources: typingmind:typingmind:training-files-vs-knowledge-base. - sync_interval — How often does an index refresh? [supported]: The KB guide describes daily connected-source refresh; Notion initial sync takes a few minutes. Neither is a permission-revocation guarantee. Sources: typingmind:typingmind:how-knowledge-base-works-in-typingmind-team, typingmind:typingmind:connect-with-your-notion-pages. - deletion_propagation — How quickly does a deleted document disappear? [unknown]: How quickly does a deleted document disappear? Deleted-document propagation and exact SaaS index/backup location were not established. - index_location — Where is indexed content stored? [partial]: Opt-in EU SaaS uses Vercel/AWS EU resources; optional self-host KB uses the operator’s PostgreSQL pgvector database and blob configuration. Exact SaaS index, backup and telemetry placement is not itemized. Sources: typingmind:typingmind:eu-data-center, typingmind:typingmind:get-started-with-self-host-installation-instruct. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:training-files-vs-knowledge-base, typingmind:typingmind:how-knowledge-base-works-in-typingmind-team, typingmind:typingmind:connect-with-your-notion-pages, typingmind:typingmind:eu-data-center, typingmind:typingmind:get-started-with-self-host-installation-instruct, typingmind:typingmind:r2-plan-matrix, typingmind:typingmind:cloud-host-vs-self-host ### KNOW-06: Grounding and retrieval quality Judgment rule: Run the shared question set over the fixture corpus: factual lookups, cross-document questions, German document with an English question, a question with no answer in the corpus. Record citation presence, link accuracy, correct answers and refusals to invent. Disclose model, settings, attempt count and failures. Single runs support a case study, not a ranking. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial The vendor describes source citations in Listener transcript search; retrieval accuracy was not tested. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - No shared corpus was loaded. Citation correctness, cross-language retrieval, refusal behaviour and observed question-set accuracy remain unmeasured. - citations — Are answers linked to specific sources? [partial]: The Listener FAQ advertises full-text search with source references over transcribed recordings. General document-RAG citation coverage is unspecified. Sources: swissgpt:swissgpt:product. - citation_accuracy — Do the cited passages support the answer? [unknown]: Do the cited passages support the answer? Unknown. No shared corpus was loaded. Citation correctness, cross-language retrieval, refusal behaviour and observed question-set accuracy remain unmeasured. - cross_language — Does a question in one language find documents in another? [unknown]: Does a question in one language find documents in another? Unknown. No shared corpus was loaded. Citation correctness, cross-language retrieval, refusal behaviour and observed question-set accuracy remain unmeasured. - no_answer_behaviour — Does it say when the corpus has no answer? [unknown]: Does it say when the corpus has no answer? Unknown. No shared corpus was loaded. Citation correctness, cross-language retrieval, refusal behaviour and observed question-set accuracy remain unmeasured. - observed_accuracy — How many of the question set were answered correctly? [unknown]: How many of the question set were answered correctly? Unknown. No shared corpus was loaded. Citation correctness, cross-language retrieval, refusal behaviour and observed question-set accuracy remain unmeasured. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:product #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Clickable source references are documented, but retrieval quality has no observed multilingual fixture. Availability: beta Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Growth/Professional SaaS RAG and reference display; model-followed instructions separate from retrieval guarantees; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Current plan matrix sells RAG in Growth/Professional, while the technical KB guide (updated2025-11-10) still labels it Early Access. No dated GA transition or tenant rollout was verified; findings retain beta/Early Access qualification. - No shared multilingual/citation-accuracy fixture. - citations — Are answers linked to specific sources? [supported]: Admins may enable reference display for knowledge responses so users can follow source links. Sources: typingmind:typingmind:get-started-with-knowledge-base. - citation_accuracy — Do the cited passages support the answer? [unknown]: Do the cited passages support the answer? No shared German/French/English question set, source-support check, cross-language retrieval test or measured answer count was run. - cross_language — Does a question in one language find documents in another? [unknown]: Does a question in one language find documents in another? No shared German/French/English question set, source-support check, cross-language retrieval test or measured answer count was run. - no_answer_behaviour — Does it say when the corpus has no answer? [partial]: Admins can instruct the model to refrain from answering when knowledge is insufficient; the system-prompt guide gives an example. This is prompt configuration, not a tested no-answer guarantee. Sources: typingmind:typingmind:system-prompts-on-admin-panel. - observed_accuracy — How many of the question set were answered correctly? [unknown]: How many of the question set were answered correctly? No shared German/French/English question set, source-support check, cross-language retrieval test or measured answer count was run. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:get-started-with-knowledge-base, typingmind:typingmind:system-prompts-on-admin-panel, typingmind:typingmind:r2-plan-matrix ### KNOW-07: Memory and personal context Judgment rule: Record personal custom instructions, cross-chat memory, project or workspace context, and user controls to view, edit, delete or disable memory. Record whether admins can disable memory organisation-wide. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Persistent agent and global admin instructions are documented; named-tier cross-chat memory and its controls remain unknown. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams SaaS Starter/Growth/Professional instructions and shared agents; exact commercial entitlement for non-project memory and its controls unverified; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Persistent instructions are distinct from learned cross-chat memory. The Memory guide and release do not establish its named-tier entitlement. - custom_instructions — Can users set persistent instructions? [supported]: Persistent agent instructions and global admin instructions are documented; agents can include full-context training files. Sources: typingmind:typingmind:system-prompts-on-admin-panel, typingmind:typingmind:training-files-vs-knowledge-base. - cross_chat_memory — Does the product remember facts across chats? [unknown]: The October 14 Teams release and Memory guide describe cross-chat memory, but no named Starter/Growth/Professional entitlement for that capability was established. The Growth/Professional Projects gate does not establish non-project memory availability. - user_memory_controls — Can users view, edit and delete memories? [unknown]: The Memory guide describes saved-memory settings, but their named Teams tier entitlement and complete edit/delete behavior remain unverified. No personal-app control is inferred to be a Teams entitlement. - admin_memory_controls — Can admins turn memory off? [unknown]: The October 14 Teams release describes an admin Memory switch, but the current plan matrix does not map it to a named tier. Organization-wide memory-control entitlement remains unknown. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:system-prompts-on-admin-panel, typingmind:typingmind:training-files-vs-knowledge-base, typingmind:typingmind:release, typingmind:typingmind:memory, typingmind:typingmind:r2-plan-matrix ## Assistants and agents ### AGT-01: Custom assistants Judgment rule: Record creation of assistants with instructions, model, knowledge and tools; who can create; sharing scope; versioning; and an internal catalogue or directory. Build one HR policy assistant on the fixture and share it with one group. Prompt templates and a prompt library are recorded here. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial Users can create document-backed assistants and use prompt templates. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - Creator restrictions, sharing boundaries, versioning, tool configuration and an organisation-managed directory are not established in the public pages. - create_assistants — Can users create assistants with instructions, knowledge and tools? [partial]: Users create workspaces with their own reference documents; configurable instructions and external tools are not described. Sources: swissgpt:swissgpt:product. - creator_permissions — Can admins limit who creates them? [unknown]: Can admins limit who creates them? Unknown. Creator restrictions, sharing boundaries, versioning, tool configuration and an organisation-managed directory are not established in the public pages. - sharing_scope — Can assistants be shared with users, groups or everyone? [unknown]: Can assistants be shared with users, groups or everyone? Unknown. Creator restrictions, sharing boundaries, versioning, tool configuration and an organisation-managed directory are not established in the public pages. - catalogue — Is there an internal assistant directory? [partial]: Premium adds preconfigured assistants; an organisation directory is not established. Sources: swissgpt:swissgpt:product. - prompt_library — Are shared prompt templates available? [partial]: Chat includes prompt templates for recurring tasks; shared-team template administration is not specified. Sources: swissgpt:swissgpt:product. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:product #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: supported Teams admins publish agents and prompts, with delegated creation permissions and group access. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Starter/Growth/Professional shared agents and prompts; Growth/Professional RAG; Professional delegated roles and resource visibility; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. - create_assistants — Can users create assistants with instructions, knowledge and tools? [supported]: All SaaS tiers include shared agents and configured models/plugins; tagged RAG knowledge requires Growth/Professional. Sources: typingmind:typingmind:build-a-shared-ai-agent-collection, typingmind:typingmind:training-files-vs-knowledge-base, typingmind:typingmind:r2-plan-matrix. - creator_permissions — Can admins limit who creates them? [supported]: Professional custom roles delegate AI Agent management separately from other Admin Panel permissions. Sources: typingmind:typingmind:roles-and-permissions, typingmind:typingmind:r2-plan-matrix. - sharing_scope — Can assistants be shared with users, groups or everyone? [supported]: All tiers publish shared agent collections; Professional adds per-user/group resource restrictions. Sources: typingmind:typingmind:build-a-shared-ai-agent-collection, typingmind:typingmind:restrict-ai-agents-access, typingmind:typingmind:r2-plan-matrix. - catalogue — Is there an internal assistant directory? [supported]: The shared agent collection is displayed in the team chat UI. Sources: typingmind:typingmind:build-a-shared-ai-agent-collection. - prompt_library — Are shared prompt templates available? [supported]: All SaaS tiers provide a shared prompt library; Professional adds per-user prompt visibility controls. Sources: typingmind:typingmind:build-a-shared-prompt-library, typingmind:typingmind:r2-plan-matrix. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:build-a-shared-ai-agent-collection, typingmind:typingmind:training-files-vs-knowledge-base, typingmind:typingmind:r2-plan-matrix, typingmind:typingmind:roles-and-permissions, typingmind:typingmind:restrict-ai-agents-access, typingmind:typingmind:build-a-shared-prompt-library ### AGT-02: Tools, actions and MCP Judgment rule: Record built-in tools, custom actions (OpenAPI), MCP client support (local and remote servers), authentication model per tool (shared service account or per-user OAuth), and admin approval of tools. Verify one external-tool call and its access boundary. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Built-in plugins and custom HTTP/JavaScript/MCP tools are available; authentication varies by connector. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel, TypingMind personal browser app and PWA; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams Starter/Growth/Professional plugins and remote MCP; Professional group restrictions; personal Private MCP Connector route separate; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=not specified; region=Remote server/provider regions depend on configuration; personal local MCP is not a verified Teams local route; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - builtin_tools — Which tools are built in? [supported]: Documented tools include Toolbox, web/deep research, image generation and voice services; optional third-party API keys may be required. Sources: typingmind:typingmind:toolbox, typingmind:typingmind:deep-research, typingmind:typingmind:image, typingmind:typingmind:voice. - custom_actions — Can admins add API actions? [supported]: Custom actions can be implemented as JavaScript, HTTP actions or MCP integrations. Sources: typingmind:typingmind:overview. - mcp_client — Can it use remote MCP servers, and with what admin control? [partial]: Teams pricing lists MCP integration and the MCP guide explicitly gives Team OAuth callbacks and per-user/per-chat sessions for remote HTTP MCP. The Private MCP Connector and local terminal-command procedure are explicitly personal-app routes; Teams local MCP support is unverified. Sources: typingmind:typingmind:mcp, typingmind:typingmind:pricing, typingmind:typingmind:r2-plan-matrix. - tool_auth — Do tools act as the user or as a shared account? [partial]: Teams remote MCP supports user OAuth and per-user sessions; separately configured plugins may use shared keys. The personal Private MCP Connector is not claimed for Teams. Sources: typingmind:typingmind:mcp, typingmind:typingmind:deep-research, typingmind:typingmind:r2-plan-matrix. - tool_approval — Can admins approve or block tools, and can users confirm actions before they run? [partial]: Professional admins restrict plugin visibility/usage by group. The MCP guide allows tools to be disabled individually; a mandatory per-action side-effect confirmation policy and Teams local connector are unverified. Sources: typingmind:typingmind:restrict-plugin-access, typingmind:typingmind:mcp, typingmind:typingmind:r2-plan-matrix. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:toolbox, typingmind:typingmind:deep-research, typingmind:typingmind:image, typingmind:typingmind:voice, typingmind:typingmind:overview, typingmind:typingmind:mcp, typingmind:typingmind:pricing, typingmind:typingmind:r2-plan-matrix, typingmind:typingmind:restrict-plugin-access ### AGT-03: Web search and deep research Judgment rule: Record web search on/off control per organisation and group, search provider, citation of web sources, deep or multi-step research, and whether queries leave the residency boundary. Admins of regulated companies often need web search off. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial Web search is included but requires organisation activation and user consent; queries can leave AlpineAI infrastructure. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - Group-specific controls, a multi-step research/report mode and general web citation behaviour are not documented. - web_search — Is web search available? [supported]: Included across licence models and optional at use time. Sources: swissgpt:swissgpt:product, swissgpt:swissgpt:terms. - admin_toggle — Can admins disable it per organisation or group? [supported]: Administrator must enable the instance; can disable it instance-wide. Users consent per search and again for detected sensitive content. Sources: swissgpt:swissgpt:terms. - search_provider — Which search provider handles queries, and where? [supported]: DPA Annex 3 names Brave Software, Inc., US, for Brave Search API; activated search sends query/chat content outside AlpineAI infrastructure. Sources: swissgpt:swissgpt:dpa, swissgpt:swissgpt:terms. - deep_research — Is multi-step research with a report available? [unknown]: Is multi-step research with a report available? Unknown. Group-specific controls, a multi-step research/report mode and general web citation behaviour are not documented. - web_citations — Are web sources cited? [unknown]: Are web sources cited? Unknown. Group-specific controls, a multi-step research/report mode and general web citation behaviour are not documented. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:product, swissgpt:swissgpt:terms, swissgpt:swissgpt:dpa #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Deep Research is a configurable search/read/report plugin with separate provider accounts. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams plugin access on Starter/Growth/Professional; Professional for group availability/limits; add_ons=Customer SerpAPI and Firecrawl keys; model API charges; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - web_search — Is web search available? [supported]: The Deep Research plugin performs iterative web search and page reading. Sources: typingmind:typingmind:deep-research. - admin_toggle — Can admins disable it per organisation or group? [supported]: Professional admins restrict the research/search plugin to selected user groups through Plugin Access and Limits. Sources: typingmind:typingmind:restrict-plugin-access, typingmind:typingmind:r2-plan-matrix. - search_provider — Which search provider handles queries, and where? [partial]: This plugin requires SerpAPI for search and Firecrawl for page retrieval; processing countries for queries/content were not established. Sources: typingmind:typingmind:deep-research. - deep_research — Is multi-step research with a report available? [supported]: The plugin plans and iterates search/reading into a report, within its documented two-minute execution bound. Sources: typingmind:typingmind:deep-research. - web_citations — Are web sources cited? [partial]: The research plugin produces a sourced report, but link completeness and citation accuracy were not independently checked. Sources: typingmind:typingmind:deep-research. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:deep-research, typingmind:typingmind:restrict-plugin-access, typingmind:typingmind:r2-plan-matrix ### AGT-04: Workflows and automation Judgment rule: Record no-code or low-code workflows, scheduled prompts, triggers from email, forms or business systems, and multi-step chains. Name where they run and who can build them. A shared prompt template is not a workflow. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial The documented automation layer uses agents and plugins rather than an established visual scheduling system. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel, TypingMind Teams REST API; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams plugins/agents on Starter/Growth/Professional; Professional external REST API integration; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - workflow_builder — Is there a workflow or chain builder? [partial]: Docs describe composing AI interactions with JavaScript/HTTP/MCP plugins and business automation examples. A visual DAG/chain builder was not established. Sources: typingmind:typingmind:overview. - scheduled_tasks — Can prompts or assistants run on a schedule? [unknown]: Can prompts or assistants run on a schedule? No native recurring scheduler, durable workflow engine or supported event trigger setup was established. - event_triggers — Can external events start a workflow? [partial]: Professional external REST integration lets an external application call a Teams chat/agent API. This is a developer invocation route, not an established native managed trigger engine. Sources: typingmind:typingmind:api-integration-overview, typingmind:typingmind:r2-plan-matrix. - builder_audience — Is building aimed at business users, IT or developers? [supported]: Team users use shared agents; technical builders configure JavaScript/HTTP plugins and credentials. Professional custom roles delegate administration, and its API supports developer integrations. Sources: typingmind:typingmind:overview, typingmind:typingmind:roles-and-permissions, typingmind:typingmind:r2-plan-matrix. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:overview, typingmind:typingmind:api-integration-overview, typingmind:typingmind:r2-plan-matrix, typingmind:typingmind:roles-and-permissions ### AGT-05: Agentic task execution Judgment rule: Record an agent mode that plans and carries out multi-step tasks with tools (browser, files, connectors), with visible steps, user confirmation of side effects and stop control. Run one fixture task (gather three documents, draft a summary, propose an email). Record interventions and what ran. A final claim of success is not proof. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Multi-step tool workflows exist; universal side-effect confirmation and run controls remain unproven. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=TypingMind Teams SaaS; Starter, Growth and Professional prerequisites stated per answer; Business quote not obtained; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - agent_mode — Is there an agent mode for multi-step tasks? [supported]: The Deep Research plugin repeatedly searches/reads, and automation docs describe tool-driven multistep business actions. Sources: typingmind:typingmind:deep-research, typingmind:typingmind:overview. - visible_steps — Are steps and tool calls shown? [partial]: The research workflow is described as iterative search and page reading, but a complete visible execution trace for all agents/tools was not established. Sources: typingmind:typingmind:deep-research. - side_effect_confirmation — Must the user confirm actions with side effects? [unknown]: Must the user confirm actions with side effects? No bounded run verified visible traces, stop behaviour, rollback or mandatory side-effect confirmation for all custom tools. - stop_control — Can the user stop a running task? [unknown]: Can the user stop a running task? No bounded run verified visible traces, stop behaviour, rollback or mandatory side-effect confirmation for all custom tools. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:deep-research, typingmind:typingmind:overview ### AGT-06: Programmable access Judgment rule: Record a public API for chat, assistants and knowledge, SDKs, API authentication, rate limits and whether API use counts against the same governance (audit, budgets, model restrictions). Being an MCP client is different from exposing the product over MCP. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial A separate AlpineAI API platform is advertised. Custom application integrations are explicitly offered under Enterprise from 10 licences; standalone API entitlement remains unknown and new accounts are activated gradually. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI API platform; version=not specified; plan=Enterprise custom application integrations from 10 licences; standalone API tariff/entitlement not established; add_ons=Individually contracted integration; API account activation required; deployment_mode=not specified; region=API portal advertises Swiss processing; endpoint-specific processing and contractual entitlement unverified; os=not specified. Limitations: - Enterprise custom integration is documented; a universal API entitlement across chat licence tiers is not. - No authenticated endpoint documentation, API account, rate-limit data or governance fixture was accessed. - chat_api — Is there an API for chat and assistants? [partial]: The product page describes requests/results through its separate API platform and explicitly places custom integrations in Enterprise. The public API landing page supports REST/SDK positioning but requires sign-in and gradual account activation. Specific chat/assistant endpoints, auth methods, standalone licensing and inclusion in Standard/Premium were not verified. Sources: swissgpt:swissgpt:review2-product, swissgpt:swissgpt:review2-api-access, swissgpt:swissgpt:university. - knowledge_api — Can documents be loaded through an API? [unknown]: No unauthenticated document-ingestion endpoint reference was retrieved from the API portal; naming general integrations is not proof of an ingestion API. - api_governance — Do audit, budgets and model rules apply to API use? [unknown]: The API landing page advertises monitoring of usage and costs, but scope of chat audit, budgets and model rules for API calls was not established without authenticated documentation. - mcp_server — Does the product expose itself to other agents over MCP? [unknown]: No AlpineAI-provided MCP server exposing the product to external agents was documented in the retrieved product or API landing pages. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:review2-product, swissgpt:swissgpt:review2-api-access, swissgpt:swissgpt:university #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial The Teams REST API creates chat/agent responses with governed api-users. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams REST API; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams Professional API integration; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=not specified; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - chat_api — Is there an API for chat and assistants? [supported]: Teams REST endpoints produce model or agent responses, including configured plugins, KB and MCP capabilities. Sources: typingmind:typingmind:api-integration-overview. - knowledge_api — Can documents be loaded through an API? [unknown]: Can documents be loaded through an API? The retrieved API overview did not establish a supported knowledge-upload API or an MCP server exposing TypingMind itself. - api_governance — Do audit, budgets and model rules apply to API use? [partial]: An API key creates an api-user subject to the same user restrictions and usage limits. Complete audit coverage, currency budgets and all API routes were not established. Sources: typingmind:typingmind:api-integration-overview. - mcp_server — Does the product expose itself to other agents over MCP? [unknown]: Does the product expose itself to other agents over MCP? The retrieved API overview did not establish a supported knowledge-upload API or an MCP server exposing TypingMind itself. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:api-integration-overview ## Daily work ### WORK-01: Clients and access points Judgment rule: List web, desktop (Windows, macOS), iOS, Android, mobile browser and browser extension separately. For each, record login with company SSO, feature gaps against web, and MDM/Intune deployability. An installable app is different from a mobile website. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial The FAQ lists web and iOS/Android access. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client, AlpineAI iOS and Android apps; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland; mobile clients on user devices; os=iOS, Android. Limitations: - Desktop OS support is only package-dependent in AGB 2.1. Browser extension, MDM/Intune deployment and mobile parity are not documented. - web — Is a full web client available? [supported]: Works in Chrome, Firefox, Safari or Edge on internet-connected computers and mobile devices. Sources: swissgpt:swissgpt:product. - desktop_apps — Windows and macOS apps? [unknown]: Windows and macOS apps? Unknown. Desktop OS support is only package-dependent in AGB 2.1. Browser extension, MDM/Intune deployment and mobile parity are not documented. - mobile_apps — iOS and Android apps with SSO? [partial]: iOS and Android apps are advertised; company SSO in those apps was not verified. Sources: swissgpt:swissgpt:product. - browser_extension — Is there a browser extension or sidebar? [unknown]: Is there a browser extension or sidebar? Unknown. Desktop OS support is only package-dependent in AGB 2.1. Browser extension, MDM/Intune deployment and mobile parity are not documented. - managed_deployment — Can IT deploy and configure clients through MDM? [unknown]: Can IT deploy and configure clients through MDM? Unknown. Desktop OS support is only package-dependent in AGB 2.1. Browser extension, MDM/Intune deployment and mobile parity are not documented. - client_feature_gaps — Which features are missing on mobile? [unknown]: Which features are missing on mobile? Unknown. Desktop OS support is only package-dependent in AGB 2.1. Browser extension, MDM/Intune deployment and mobile parity are not documented. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:product #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial The primary client is a browser app; installable PWAs are documented separately from native apps. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel, TypingMind personal browser app and PWA; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams SaaS Starter/Growth/Professional browser chat; PWA installation findings only for personal typingmind.com (licence feature gates separate); add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=not specified; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - web — Is a full web client available? [supported]: Teams supplies a branded browser chat interface and separate Admin Panel. Sources: typingmind:typingmind:typingmind-vs-typingmind-team. - desktop_apps — Windows and macOS apps? [partial]: The personal typingmind.com app has browser-PWA installation instructions for Windows/macOS; this is not evidence of a Teams native desktop app or Teams PWA deployment entitlement. Sources: typingmind:typingmind:install, typingmind:typingmind:r2-personal-bulk-licence. - mobile_apps — iOS and Android apps with SSO? [partial]: Personal typingmind.com can be added to iOS/Android home screens as a browser PWA. Teams Professional SSO is a separate entitlement; no Teams installable mobile client or mobile SSO fixture was verified. Sources: typingmind:typingmind:install, typingmind:typingmind:r2-personal-bulk-licence, typingmind:typingmind:r2-plan-matrix. - browser_extension — Is there a browser extension or sidebar? [unknown]: Is there a browser extension or sidebar? No MDM packaging/configuration, supported browser sidebar/extension or mobile feature-parity matrix was established. - managed_deployment — Can IT deploy and configure clients through MDM? [unknown]: Can IT deploy and configure clients through MDM? No MDM packaging/configuration, supported browser sidebar/extension or mobile feature-parity matrix was established. - client_feature_gaps — Which features are missing on mobile? [unknown]: Which features are missing on mobile? No MDM packaging/configuration, supported browser sidebar/extension or mobile feature-parity matrix was established. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:typingmind-vs-typingmind-team, typingmind:typingmind:install, typingmind:typingmind:r2-personal-bulk-licence, typingmind:typingmind:r2-plan-matrix ### WORK-02: Office suite and Teams integration Judgment rule: Record presence inside Word, Excel, PowerPoint, Outlook and Teams (and Google Docs, Gmail), with what the assistant can do in each: read the open document, write into it, act on mail. Record extra licences and admin consent needed. A connector that searches Office files is KNOW-03, not WORK-02. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial Standard and Premium include an Outlook add-in for reading, drafting and translating email. Availability: generally-available Delivery: official-integration Scope: surfaces=AlpineAI Outlook add-in; version=not specified; plan=Standard and Premium; Enterprise packaging unspecified; add_ons=none stated; deployment_mode=not specified; region=Content processed on AlpineAI hardware in Switzerland; os=not specified. Limitations: - Word, Excel, PowerPoint, Teams and Google application integrations are not specified. Outlook add-in consent and compatibility requirements remain unverified. - word_excel_powerpoint — Does it work inside Word, Excel and PowerPoint? [unknown]: Does it work inside Word, Excel and PowerPoint? Unknown. Word, Excel, PowerPoint, Teams and Google application integrations are not specified. Outlook add-in consent and compatibility requirements remain unverified. - outlook — Does it work inside Outlook? [supported]: Add-in summarises messages/threads, drafts replies and translates within Outlook. The FAQ describes Get Add-Ins installation. Sources: swissgpt:swissgpt:product. - teams — Is there a Teams app or bot? [unknown]: Is there a Teams app or bot? Unknown. Word, Excel, PowerPoint, Teams and Google application integrations are not specified. Outlook add-in consent and compatibility requirements remain unverified. - google_workspace_apps — Does it work inside Docs, Sheets or Gmail? [unknown]: Does it work inside Docs, Sheets or Gmail? Unknown. Word, Excel, PowerPoint, Teams and Google application integrations are not specified. Outlook add-in consent and compatibility requirements remain unverified. - licence_prerequisites — Which licences or consents are needed? [partial]: Included with Standard and Premium. Microsoft licence/version prerequisites and tenant consent requirements are not stated. Sources: swissgpt:swissgpt:product. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:product #### TypingMind / TypingMind Teams Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. ### WORK-03: Translation and writing Judgment rule: Record document translation preserving layout, the translation engine (LLM or dedicated such as DeepL), glossaries, formality control, supported languages including Swiss German conventions (no ß), and rewriting/proofreading tools. Translate the shared German contract into French and Italian and record layout and glossary results. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial The platform and Outlook add-in offer drafting, summaries and translations. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - Whole-document layout preservation, a dedicated translation engine, company glossaries, formality and de-CH controls were not specified or tested. - document_translation — Can whole documents be translated with formatting kept? [unknown]: Can whole documents be translated with formatting kept? Unknown. Whole-document layout preservation, a dedicated translation engine, company glossaries, formality and de-CH controls were not specified or tested. - translation_engine — Which engine translates? [unknown]: Which engine translates? Unknown. Whole-document layout preservation, a dedicated translation engine, company glossaries, formality and de-CH controls were not specified or tested. - glossaries — Are company glossaries supported? [unknown]: Are company glossaries supported? Unknown. Whole-document layout preservation, a dedicated translation engine, company glossaries, formality and de-CH controls were not specified or tested. - formality_and_variants — Can formality and regional variants (de-CH) be set? [unknown]: Can formality and regional variants (de-CH) be set? Unknown. Whole-document layout preservation, a dedicated translation engine, company glossaries, formality and de-CH controls were not specified or tested. - writing_tools — Are dedicated rewrite and proofreading tools available? [partial]: The FAQ advertises drafting and replying to email, summaries and translations into Swiss national languages and English. Proofreading controls are not detailed. Sources: swissgpt:swissgpt:product. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:product #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Configurable model instructions support writing workflows; dedicated translation fidelity is unestablished. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=TypingMind Teams SaaS; Starter, Growth and Professional prerequisites stated per answer; Business quote not obtained; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - document_translation — Can whole documents be translated with formatting kept? [unknown]: Can whole documents be translated with formatting kept? No whole-document formatting-preserving translation fixture, company glossary engine or Swiss language-quality test was retrieved. - translation_engine — Which engine translates? [partial]: Writing/translation prompts are handled by the customer-selected LLM; no separate translation engine or formatting-preserving document pipeline was established. Sources: typingmind:typingmind:manage-ai-models, typingmind:typingmind:system-prompts-on-admin-panel. - glossaries — Are company glossaries supported? [unknown]: Are company glossaries supported? No whole-document formatting-preserving translation fixture, company glossary engine or Swiss language-quality test was retrieved. - formality_and_variants — Can formality and regional variants (de-CH) be set? [partial]: Global and agent instructions can specify tone and language, but no dedicated de-CH/formality switch or quality guarantee was established. Sources: typingmind:typingmind:system-prompts-on-admin-panel. - writing_tools — Are dedicated rewrite and proofreading tools available? [partial]: Shared prompts and marketing-writing workflows support rewriting tasks; a dedicated proofreading interface and correction-quality fixture were not established. Sources: typingmind:typingmind:build-a-shared-prompt-library, typingmind:typingmind:ad-creative-generator. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:manage-ai-models, typingmind:typingmind:system-prompts-on-admin-panel, typingmind:typingmind:build-a-shared-prompt-library, typingmind:typingmind:ad-creative-generator ### WORK-04: Image generation and editing Judgment rule: Record available image models, editing (inpainting, reference images), admin control and per-image cost visibility, and licence or usage terms for generated images. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial GPT Image generation and editing use customer API credentials and controllable plugins. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams plugin access on Starter/Growth/Professional; Professional for group-based image-plugin governance; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - image_generation — Can users generate images, with which models? [supported]: The GPT Image 2.5 plugin documents image generation through the customer’s OpenAI API. Sources: typingmind:typingmind:image. - image_editing — Can users edit uploaded images? [supported]: The October 2 changelog says GPT Image editing accepts any image in a chat, including attachments and artifact images. Sources: typingmind:typingmind:release, typingmind:typingmind:image. - admin_control — Can admins restrict image generation? [partial]: Professional admins restrict the image plugin by user group. Coverage of every provider-native image route remains unverified. Sources: typingmind:typingmind:restrict-plugin-access, typingmind:typingmind:r2-plan-matrix. - usage_terms — What usage rights apply to output? [partial]: The Terms say users own their content, subject to provider terms. No independent clearance of third-party rights or all model-provider output licences was performed. Sources: typingmind:typingmind:terms-of-service. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:image, typingmind:typingmind:release, typingmind:typingmind:restrict-plugin-access, typingmind:typingmind:r2-plan-matrix, typingmind:typingmind:terms-of-service ### WORK-05: Voice and meetings Judgment rule: Record dictation, read-aloud, live voice conversation, meeting transcription and summarisation, supported languages including Swiss German speech, and where audio is processed. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial Premium Listener records audio, transcribes Swiss German and creates minutes or summaries. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client, AlpineAI iOS and Android apps; version=not specified; plan=Premium Listener; Enterprise inclusion per order; add_ons=none stated; deployment_mode=not specified; region=Audio processing on AlpineAI-owned hardware in Switzerland; os=not specified. Limitations: - Read-aloud and live bidirectional voice conversation are undocumented; recognition quality was not independently tested. - dictation — Can users dictate prompts? [partial]: Listener records dictation through the microphone and accepts existing audio files. Direct dictation into a chat prompt was not established. Sources: swissgpt:swissgpt:product. - read_aloud — Can answers be read aloud? [unknown]: Can answers be read aloud? Unknown. Read-aloud and live bidirectional voice conversation are undocumented; recognition quality was not independently tested. - voice_conversation — Is a live voice mode available? [unknown]: Is a live voice mode available? Unknown. Read-aloud and live bidirectional voice conversation are undocumented; recognition quality was not independently tested. - meeting_summaries — Are meetings transcribed and summarised? [supported]: Advertises complete transcripts and structured minutes, summaries, key facts and to-dos; recordings process on AlpineAI hardware in Switzerland. Sources: swissgpt:swissgpt:product. - swiss_german_speech — Is Swiss German speech recognised? [supported]: Vendor claims Swiss German and other dialect recognition, with transcript output in German, English, French or Italian. Sources: swissgpt:swissgpt:product. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:product #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Dictation, read-aloud and audio-summary workflows are documented; live meetings and Swiss German quality are unknown. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=TypingMind Teams SaaS; Starter, Growth and Professional prerequisites stated per answer; Business quote not obtained; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - dictation — Can users dictate prompts? [supported]: Dictation uses browser speech recognition or customer OpenAI Whisper, Groq or Azure transcription credentials; Teams admins configure keys. Sources: typingmind:typingmind:voice. - read_aloud — Can answers be read aloud? [supported]: Browser speech, ElevenLabs or OpenAI TTS reads answers aloud with playback/autoplay controls. Sources: typingmind:typingmind:tts. - voice_conversation — Is a live voice mode available? [unknown]: Is a live voice mode available? No full-duplex live voice mode, Swiss German audio fixture or native meeting-attendance integration was established. - meeting_summaries — Are meetings transcribed and summarised? [partial]: The meeting-summarizer workflow uses recorded audio transcription and LLM summaries; Whisper upload is capped at 25MB. No live meeting bot or diarization quality test was established. Sources: typingmind:typingmind:meeting-summarizer, typingmind:typingmind:voice. - swiss_german_speech — Is Swiss German speech recognised? [unknown]: Is Swiss German speech recognised? No full-duplex live voice mode, Swiss German audio fixture or native meeting-attendance integration was established. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:voice, typingmind:typingmind:tts, typingmind:typingmind:meeting-summarizer ### WORK-06: Data analysis and document output Judgment rule: Record code execution for data analysis, charts, and generation of downloadable Word, Excel, PowerPoint and PDF files. Analyse the fixture Excel and request a PowerPoint summary. Record where code runs and whether files are editable. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: supported Toolbox executes code locally and produces charts and Office documents; Growth includes Canvas. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams Growth/Professional for Canvas; Toolbox on tool-capable models with separate API usage; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. - data_analysis — Can it analyse spreadsheets with executed code? [supported]: Toolbox runs Python/JavaScript in the browser, including tabular analysis libraries. Network is disabled by default; analysis is not a vendor EU code-execution guarantee. Sources: typingmind:typingmind:toolbox. - charts — Can it produce charts? [supported]: Toolbox supports chart generation with Plotly and Bokeh. Sources: typingmind:typingmind:toolbox. - office_file_output — Can it create Word, Excel or PowerPoint files? [supported]: Toolbox creates/edits documents, spreadsheets, presentations and PDFs, including Office formats. Sources: typingmind:typingmind:toolbox. - canvas_editing — Is there a document or canvas editor for iterating on output? [supported]: The Teams plan comparison lists Canvas and Artifacts for Growth and Professional. Sources: typingmind:typingmind:compare-typingmind-team-plans. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:toolbox, typingmind:typingmind:compare-typingmind-team-plans ### WORK-07: Sharing and team spaces Judgment rule: Record chat sharing (link, user, group), shared projects or team spaces, channels or group chat with AI, and commenting. Test sharing a chat with a colleague in another group and record what the recipient sees, including knowledge they lack access to. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Shared agents and prompt libraries are documented, but collaborative project/chat boundaries are not established. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=TypingMind Teams SaaS; Starter, Growth and Professional prerequisites stated per answer; Business quote not obtained; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - share_chats — Can chats be shared inside the organisation? [unknown]: Can chats be shared inside the organisation? No Teams-specific shared-chat recipient ACL/redaction guarantee, simultaneous multi-person AI conversation or source-permission sharing fixture was established. - team_spaces — Are shared projects or workspaces available? [partial]: Teams share an agent catalogue and prompt library; project-scoped memory is documented. Shared project membership and cross-user workspace semantics were not established. Sources: typingmind:typingmind:build-a-shared-ai-agent-collection, typingmind:typingmind:build-a-shared-prompt-library, typingmind:typingmind:release. - group_chat — Can several people chat together with the AI? [unknown]: Can several people chat together with the AI? No Teams-specific shared-chat recipient ACL/redaction guarantee, simultaneous multi-person AI conversation or source-permission sharing fixture was established. - share_permission_boundary — Does sharing respect the recipient's access to sources? [unknown]: Does sharing respect the recipient's access to sources? No Teams-specific shared-chat recipient ACL/redaction guarantee, simultaneous multi-person AI conversation or source-permission sharing fixture was established. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:build-a-shared-ai-agent-collection, typingmind:typingmind:build-a-shared-prompt-library, typingmind:typingmind:release ### WORK-08: Interface languages and accessibility Judgment rule: Record UI languages, focusing on German, French, Italian and English, with translation quality observed (complete, partial, machine-translated). Record response language handling and WCAG conformance statements. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial Terms specify a German and English interface; language quality was not observed. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - No authenticated German/French/Italian UI review, WCAG conformance statement or German admin-interface specification was located. - ui_languages — Which UI languages are offered? [partial]: AGB 2.1 explicitly names German and English UI. Sector pages advertise processing of Swiss national languages and English; that does not establish French/Italian UI coverage. Sources: swissgpt:swissgpt:terms. - dach_language_quality — Are German, French and Italian UIs complete and natural? [unknown]: Are German, French and Italian UIs complete and natural? Unknown. No authenticated German/French/Italian UI review, WCAG conformance statement or German admin-interface specification was located. - accessibility_statement — Is there a WCAG conformance statement? [unknown]: Is there a WCAG conformance statement? Unknown. No authenticated German/French/Italian UI review, WCAG conformance statement or German admin-interface specification was located. - admin_ui_languages — Is the admin interface available in German? [unknown]: Is the admin interface available in German? Unknown. No authenticated German/French/Italian UI review, WCAG conformance statement or German admin-interface specification was located. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:terms #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Multilingual capability is advertised without a verified DACH UI or accessibility matrix. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=TypingMind Teams SaaS; Starter, Growth and Professional prerequisites stated per answer; Business quote not obtained; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - ui_languages — Which UI languages are offered? [partial]: Teams plan documentation advertises a multilingual chat interface and a public translation collaboration project exists. A complete current language list was not established. Sources: typingmind:typingmind:compare-typingmind-team-plans, typingmind:typingmind:translations. - dach_language_quality — Are German, French and Italian UIs complete and natural? [unknown]: Are German, French and Italian UIs complete and natural? No complete German/French/Italian UI review, German Admin Panel guarantee, WCAG conformance statement or assistive-technology fixture was retrieved. - accessibility_statement — Is there a WCAG conformance statement? [unknown]: Is there a WCAG conformance statement? No complete German/French/Italian UI review, German Admin Panel guarantee, WCAG conformance statement or assistive-technology fixture was retrieved. - admin_ui_languages — Is the admin interface available in German? [unknown]: Is the admin interface available in German? No complete German/French/Italian UI review, German Admin Panel guarantee, WCAG conformance statement or assistive-technology fixture was retrieved. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:compare-typingmind-team-plans, typingmind:typingmind:translations ## Administration and governance ### ADM-01: Single sign-on and authentication Judgment rule: Record SAML, OIDC, Microsoft Entra ID, Google, LDAP, MFA enforcement, conditional-access compatibility, domain verification and plan restrictions. Test one SSO login on a trial where possible. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Teams Professional provides SAML SSO and OAuth/OIDC options. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams Professional identity features; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - sso_protocols — Which SSO protocols and identity providers work? [supported]: SAML SSO accepts IdP metadata and multiple connections; OAuth/OIDC includes Microsoft, Google and custom provider configuration. Sources: typingmind:typingmind:sso, typingmind:typingmind:o-auth-2-0. - sso_plan — Which plan includes SSO? [supported]: The plan comparison places SSO/OAuth and Directory Sync in Professional. Sources: typingmind:typingmind:compare-typingmind-team-plans. - mfa — Can MFA be enforced? [partial]: The DPA security measures mention MFA and SSO; documented tenant-level enforcement settings and IdP MFA propagation were not verified. Sources: typingmind:typingmind:data-processing-agreement. - conditional_access — Do Entra conditional-access policies apply? [unknown]: Do Entra conditional-access policies apply? No Entra conditional-access test or verified-domain capture/mandatory login policy was established. - domain_capture — Can the company claim its email domain? [unknown]: Can the company claim its email domain? No Entra conditional-access test or verified-domain capture/mandatory login policy was established. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:sso, typingmind:typingmind:o-auth-2-0, typingmind:typingmind:compare-typingmind-team-plans, typingmind:typingmind:data-processing-agreement ### ADM-02: User provisioning and lifecycle Judgment rule: Record SCIM, just-in-time provisioning, group sync from the identity provider, deprovisioning effects (chats, assistants, knowledge ownership transfer) and licence assignment. Deprovision a test user and record what happens to their content. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Professional supports SCIM, group sync and JIT; OAuth revocation alone does not remove a seat. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams Professional Directory Sync; distinguish SCIM suspension from OAuth revocation; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - scim — Is SCIM provisioning supported? [supported]: SCIM v2 creates, updates and suspends users from a compatible IdP. Sources: typingmind:typingmind:scim. - group_sync — Are identity-provider groups synced? [supported]: Directory Sync assigns users to corresponding groups received from the identity provider. Sources: typingmind:typingmind:scim. - jit_provisioning — Are users created at first login? [supported]: OAuth login can create a user on first sign-in and consumes an available seat; creation fails without a free seat. Sources: typingmind:typingmind:o-auth-2-0. - deprovisioning — What happens to a removed user's data and shared assets? [partial]: SCIM suspends users; separate OAuth docs warn that revoking IdP authorization alone does not remove an existing user, requiring admin removal. Shared asset ownership, deletion and session-revocation timing are not established. Sources: typingmind:typingmind:scim, typingmind:typingmind:o-auth-2-0. - seat_assignment — Can admins assign and reclaim seats? [partial]: Admins invite/manage users and JIT consumes purchased seats. Exact seat reclamation timing, licence proration and asset transfer were not established. Sources: typingmind:typingmind:invite-team-members, typingmind:typingmind:o-auth-2-0. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:scim, typingmind:typingmind:o-auth-2-0, typingmind:typingmind:invite-team-members ### ADM-03: Roles, groups and feature permissions Judgment rule: Record admin roles, delegated admins, groups, per-feature permissions (web search, image generation, connectors, file upload, agents), separation of subsidiaries or departments, and enforcement. Test one disabled feature for one group. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Custom admin roles and user groups provide delegated management and feature availability controls. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams Professional custom roles, resource groups and usage limits; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - admin_roles — Which admin roles exist, and can admin rights be delegated? [supported]: Default Admin has highest permissions. Owners create custom roles for models, plugins, agents, reporting, billing, users, keys, KB and portal settings. Users may hold multiple roles; the former Content Moderator preset was removed. Sources: typingmind:typingmind:roles-and-permissions. - groups — Can permissions be set per group? [supported]: User groups govern resource visibility and limits. Admin roles are assigned to individual members; assigning roles to groups is still described as coming soon. Sources: typingmind:typingmind:roles-and-permissions, typingmind:typingmind:create-and-manage-user-groups. - feature_toggles — Which features can be switched on or off per group? [supported]: Professional provides model, agent and plugin group availability/usage controls through resource Access/Limit settings; the plan matrix separately lists per-user prompt visibility. CSS/themes are excluded from this permission finding. Sources: typingmind:typingmind:restrict-model-access, typingmind:typingmind:restrict-ai-agents-access, typingmind:typingmind:restrict-plugin-access, typingmind:typingmind:r2-plan-matrix. - multi_entity — Can subsidiaries or departments be separated within one contract? [unknown]: Can subsidiaries or departments be separated within one contract? No multi-subsidiary legal/entity separation, independent subtenant keys or consolidated contract administration was established. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:roles-and-permissions, typingmind:typingmind:create-and-manage-user-groups, typingmind:typingmind:restrict-model-access, typingmind:typingmind:restrict-ai-agents-access, typingmind:typingmind:restrict-plugin-access, typingmind:typingmind:r2-plan-matrix ### ADM-04: Usage and adoption analytics Judgment rule: Record dashboards for active users, usage by feature, model, group and assistant, export, and privacy controls (aggregation, pseudonymisation, works-council-friendly settings). Distinguish aggregate adoption metrics from per-person monitoring. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Teams analytics reports adoption and consumption, including identifiable user breakdowns. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams Professional analytics; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - adoption_dashboard — Is there an adoption dashboard? [supported]: Dashboard counts active users, messages, conversations, characters and estimated tokens over a selected period. Sources: typingmind:typingmind:chatbot-analytics. - breakdowns — Can usage be broken down by group, model, feature and assistant? [supported]: Reporting filters by user, group, model and agent; a complete per-feature cost map was not established. Sources: typingmind:typingmind:chatbot-analytics. - export — Can usage data be exported? [partial]: Charts can be downloaded; a complete raw usage ledger export format was not established. Sources: typingmind:typingmind:chatbot-analytics. - privacy_modes — Can per-person reporting be limited or pseudonymised? [unknown]: Can per-person reporting be limited or pseudonymised? No pseudonymised reporting mode or enforceable per-person aggregation threshold was established. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:chatbot-analytics ### ADM-05: Spend controls and cost allocation Judgment rule: Record cost visibility per user, group and cost centre, budgets and their behaviour (alert, block, downgrade model), pooled credits, and chargeback exports. Test a small budget on a trial where possible. Distinguish billed amounts, estimates and unavailable amounts. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Admins can inspect estimated consumption and set token/message limits; monetary budgets are unestablished. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams Professional; API charges billed directly by customer providers; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - cost_visibility — Can admins see cost or consumption per user and group? [partial]: Analytics shows estimated token and character usage per user/group/model/agent. Customer model bills remain external; reconciled currency cost is not established. Sources: typingmind:typingmind:chatbot-analytics. - budgets — Can budgets be set per user or group? [partial]: Admins configure reusable token/message/character limits for users/groups and resources, rather than an evidenced currency-denominated budget. Sources: typingmind:typingmind:usage-limits. - budget_enforcement — Does reaching a budget alert, block or downgrade? [partial]: Users are shown applicable limits, but hard monetary blocking, alerts, downgrade and in-flight overshoot semantics were not established. Sources: typingmind:typingmind:usage-limits. - chargeback — Can costs be exported per cost centre? [unknown]: Can costs be exported per cost centre? No cost-centre chargeback ledger or exact monetary budget-enforcement documentation was retrieved. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:chatbot-analytics, typingmind:typingmind:usage-limits ### ADM-06: Audit logs and compliance export Judgment rule: Record which events are logged (logins, admin changes, prompts, responses, file access, connector queries), retention, export or SIEM streaming, and eDiscovery/legal hold. Use the word audit only with a defined retention and integrity model. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Optional chat recording exposes prompts and answers to authorized admins with retention and JSON export. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams Professional chat logs; logging must be enabled; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - admin_events — Are admin and security events logged? [unknown]: Are admin and security events logged? No first-party security/admin event catalogue, legal hold or eDiscovery workflow was established. - content_logging — Can prompts and answers be retained for compliance, and is it configurable? [supported]: Record all chats is opt-in and off by default; authorized Admin Panel roles can view saved user prompts and model responses. Sources: typingmind:typingmind:view-and-export-user-chat-logs, typingmind:typingmind:export-chat-logs, typingmind:typingmind:roles-and-permissions. - siem_export — Can logs stream to a SIEM or be exported? [partial]: An API exports recorded chats as JSON, last 100 matching conversations per request with date filtering. Native SIEM streaming and complete audit-event export were not established. Sources: typingmind:typingmind:export-chat-logs. - ediscovery — Are eDiscovery and legal hold supported? [unknown]: Are eDiscovery and legal hold supported? No first-party security/admin event catalogue, legal hold or eDiscovery workflow was established. - log_retention — How long are logs kept? [partial]: Admins set chat retention and old records are permanently deleted. A default duration and separate admin/security log retention were not established. Sources: typingmind:typingmind:view-and-export-user-chat-logs. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:view-and-export-user-chat-logs, typingmind:typingmind:export-chat-logs, typingmind:typingmind:roles-and-permissions ### ADM-07: Data loss prevention and personal data Judgment rule: Record detection and handling of personal and confidential data before it reaches a model: redaction, masking, blocking, warning. Record integration with Purview, Netskope or similar, and whether filtering runs locally or at a third party. Send a fixture prompt with a Swiss AHV number and IBAN and record the outcome. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial Web search has a sensitive-content consent step; general PII/DLP handling is unverified. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - Custom patterns, external DLP integration, detection coverage and filter location are not documented; no AHV/IBAN fixture was used. - pii_detection — Is personal data detected in prompts and files? [partial]: AGB 2.5 says the system can detect potentially sensitive search-query content; it does not specify personal-data types or uploaded-file detection. Sources: swissgpt:swissgpt:terms. - pii_action — Is it redacted, masked, blocked or only warned? [partial]: Detected potentially sensitive web-search content requires renewed user consent before transmission. Redaction or blocking is not established. Sources: swissgpt:swissgpt:terms. - custom_patterns — Can admins add custom patterns or terms? [unknown]: Can admins add custom patterns or terms? Unknown. Custom patterns, external DLP integration, detection coverage and filter location are not documented; no AHV/IBAN fixture was used. - dlp_integration — Does it integrate with an existing DLP product? [unknown]: Does it integrate with an existing DLP product? Unknown. Custom patterns, external DLP integration, detection coverage and filter location are not documented; no AHV/IBAN fixture was used. - filter_location — Where does filtering run? [unknown]: Where does filtering run? Unknown. Custom patterns, external DLP integration, detection coverage and filter location are not documented; no AHV/IBAN fixture was used. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:terms #### TypingMind / TypingMind Teams Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. ### ADM-08: AI policy and guardrails Judgment rule: Record organisation-wide system instructions, acceptable-use notices and acknowledgement, content filters, topic restrictions, AI-output labelling, and admin visibility of policy violations. Separate enforced policy from instructions the model is expected to follow. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Global instructions can govern chat behaviour, but policy acceptance and violation reporting remain unknown. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams SaaS global instructions; Professional for group/resource plugin restrictions; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - org_instructions — Can admins set organisation-wide instructions? [supported]: Admins set hidden global system instructions, few-shot examples and post-message reminders across chats and agents. Sources: typingmind:typingmind:system-prompts-on-admin-panel. - usage_notice — Can users be required to accept a usage policy? [unknown]: Can users be required to accept a usage policy? No mandatory policy-acceptance record, enforcement-grade content filter or admin policy-hit report was established. - content_filters — Are content filters configurable? [partial]: Admins configure model-followed instructions; Professional separately restricts plugin availability by group. Neither establishes deterministic PII/content filtering or a policy-enforcement engine. Sources: typingmind:typingmind:system-prompts-on-admin-panel, typingmind:typingmind:restrict-plugin-access, typingmind:typingmind:r2-plan-matrix. - violation_reporting — Are policy hits visible to admins? [unknown]: Are policy hits visible to admins? No mandatory policy-acceptance record, enforcement-grade content filter or admin policy-hit report was established. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:system-prompts-on-admin-panel, typingmind:typingmind:restrict-plugin-access, typingmind:typingmind:r2-plan-matrix ## Data protection and sovereignty ### DATA-01: Hosting location and data residency Judgment rule: Record country or region separately for chat storage, files and indexes, model inference, backups, logs and telemetry, and support access. Name the cloud provider. Record whether residency is default, a paid option or plan-dependent. "EU data boundary" programmes are recorded with their documented exceptions. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial Swiss RIZ application hosting is consistently described. Green/RZO Swiss inference is a vendor security-page statement; a current contracted inference-site schedule is unverified because the July DPA does not list RZO. Storage, inference, support and optional search are scoped separately. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client, AlpineAI API platform; version=Public infrastructure page and DPA Annex 3 dated 10 July 2026; no tenant contract inspected; plan=Current AlpineAI sector offer; contract-specific residency and secret-data supplement conditional on signature; add_ons=none stated; deployment_mode=not specified; region=Advertised Swiss application hosting; current contracted inference map unverified; DPA CH/EEA with conditional third-country processing and optional external search; os=not specified. Limitations: - RZO is described as colocation, not as a foreign model API. Its absence from the July annex is not evidence of overseas inference or unauthorised processing. - The currently contracted complete inference-site list, per-index/log/telemetry locations and exhaustive staff access countries remain unknown. - storage_region — Where are chats, files and indexes stored? [partial]: The security page describes application/data storage on AlpineAI hardware at two Swiss RIZ sites. July DPA Annex 3 independently lists RIZ CH for platform hosting and support. Separate per-index/log topology and a tenant-specific deployment schedule were not established. Sources: swissgpt:swissgpt:review2-security, swissgpt:swissgpt:review2-dpa. - inference_region — Where do model calls run, per provider? [unknown]: The security page names Swiss Green and RZO GPU colocation on AlpineAI-owned/operated hardware without logical access by site operators. July DPA Annex 3 names Green and expressly describes pure colocation, but omits RZO. No newer public annex or contract-specific RZO authorisation was retrieved. The currently contracted complete inference-site map is therefore unknown; the marketing statement is not a verified current contract commitment. - backups_logs_region — Where are backups, logs and telemetry? [partial]: July DPA Annex 3 lists cloudscale.ch CH for backup; Annex 2 describes daily offsite backups. This dated schedule does not establish every current log/telemetry or backup location. Sources: swissgpt:swissgpt:review2-dpa. - support_access_region — From where can vendor staff access data? [partial]: The dated annex assigns RIZ CH to customer support and EU Atlassian/Custify to helpdesk/account functions. Physical staff access countries are not fully mapped. A separately signed secret-data supplement requires Swiss services where secret data is in plaintext. Sources: swissgpt:swissgpt:review2-dpa, swissgpt:swissgpt:secrecy. - swiss_hosting — Is Swiss hosting available? [partial]: Swiss application hosting at RIZ is supported by both public sources. The full current inference-site inventory is withheld pending a current applicable schedule, so this is not an all-processing CH-residency assurance. Sources: swissgpt:swissgpt:review2-security, swissgpt:swissgpt:review2-dpa. - residency_plan — Which plan or option provides residency? [partial]: DPA 1.5 permits CH/EEA processing with conditional third-country safeguards. Optional Brave search is separately listed with US processing. The secret-data supplement, only when signed and applicable, narrows plaintext services including APIs to Switzerland. A current tenant-specific inference schedule is still missing. Sources: swissgpt:swissgpt:review2-dpa, swissgpt:swissgpt:secrecy. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:review2-security, swissgpt:swissgpt:review2-dpa, swissgpt:swissgpt:secrecy #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial EU SaaS storage is optional; inference, tools and support are separate residency questions. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel, Commercial self-hosted TypingMind Teams server; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=TypingMind Teams SaaS; Starter, Growth and Professional prerequisites stated per answer; Business quote not obtained; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=not specified; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - storage_region — Where are chats, files and indexes stored? [partial]: Teams offers US and opt-in EU instances. The EU guide names Vercel web hosting and AWS database resources; exact file/index/backup components are not fully enumerated. Sources: typingmind:typingmind:eu-data-center, typingmind:typingmind:privacy-policy. - inference_region — Where do model calls run, per provider? [supported]: The EU guide explicitly warns model calls may go outside the EU. Customers can choose an EU Azure endpoint or private/local model; this is not an all-provider EU inference promise. Sources: typingmind:typingmind:eu-data-center. - backups_logs_region — Where are backups, logs and telemetry? [unknown]: Where are backups, logs and telemetry? Backup, telemetry, support-access country matrix and contractual Swiss-only end-to-end processing were not established. - support_access_region — From where can vendor staff access data? [partial]: The DPA identifies the processor at a Da Nang, Vietnam address and requires need-to-know confidentiality; actual operational staff access countries and restrictions were not established. Sources: typingmind:typingmind:data-processing-agreement. - swiss_hosting — Is Swiss hosting available? [partial]: Customer self-host can run on operator-selected infrastructure, but no vendor-operated Swiss SaaS region or complete CH-only processing undertaking was established. Sources: typingmind:typingmind:selfhost, typingmind:typingmind:get-started-with-self-host-installation-instruct. - residency_plan — Which plan or option provides residency? [supported]: The EU instance is an account/instance choice for Teams SaaS. Self-host is a separate quoted annual commercial licence, not the personal licence. Sources: typingmind:typingmind:eu-data-center, typingmind:typingmind:selfhost. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:eu-data-center, typingmind:typingmind:privacy-policy, typingmind:typingmind:data-processing-agreement, typingmind:typingmind:selfhost, typingmind:typingmind:get-started-with-self-host-installation-instruct ### DATA-02: Deployment options Judgment rule: Record multi-tenant SaaS, dedicated single-tenant instance, deployment in the customer's cloud account, on-premise and air-gapped as separate options with plan and availability. Record which features are missing in each mode, who operates it, and the documented seat or contract threshold for each mode, which can differ from the enterprise plan threshold (e.g. dedicated SaaS from 2,000 seats, own cloud from 5,000). A local-network connector or gateway that reaches on-prem data from a vendor-hosted platform is KNOW-03, not on-premise deployment. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Teams offers multitenant SaaS and a separate commercial self-host package; offline operation is not established. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel, Commercial self-hosted TypingMind Teams server; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=TypingMind Teams SaaS; Starter, Growth and Professional prerequisites stated per answer; Business quote not obtained; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=not specified; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Marketing language about data staying on customer infrastructure must be read with model, embedding, licence and update egress requirements. - multi_tenant_saas — Is a shared SaaS offered? [supported]: The privacy policy explicitly describes Teams as server-stored multitenant SaaS, unlike the default local-storage personal app. Sources: typingmind:typingmind:privacy-policy. - dedicated_instance — Is a dedicated single-tenant instance offered, from what size? [partial]: Self-host permits one customer-operated instance with five included users. A separate vendor-operated dedicated SaaS service and its minimum were not established. Sources: typingmind:typingmind:selfhost. - customer_cloud — Can it run in the customer's own cloud account, from what size? [supported]: A quoted self-host licence runs the NodeJS service, MySQL, optional PostgreSQL KB and blob store in the customer’s cloud account. Sources: typingmind:typingmind:selfhost, typingmind:typingmind:get-started-with-self-host-installation-instruct. - on_premise — Can the whole platform run on customer infrastructure, from what size? [supported]: The self-host package can run on customer infrastructure with Docker/NodeJS and required databases; five users and one instance are included under the annual quote. Sources: typingmind:typingmind:selfhost. - air_gapped — Can it run without internet access? [partial]: The self-host FAQ lists external model requests plus update and licence checks. Private models reduce model egress, while the optional KB guide still requires OpenAI embeddings. A supported air-gapped licence was not established. Sources: typingmind:typingmind:selfhost, typingmind:typingmind:get-started-with-self-host-installation-instruct. - mode_feature_gaps — Which features are missing outside SaaS? [supported]: Self-host releases lag cloud by approximately one to three weeks; optional KB needs additional infrastructure, external OAuth setup and OpenAI embeddings. SLA/dedicated support is a separate paid option. Sources: typingmind:typingmind:selfhost, typingmind:typingmind:setup-external-connectors-for-knowledge-base, typingmind:typingmind:get-started-with-self-host-installation-instruct. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:privacy-policy, typingmind:typingmind:selfhost, typingmind:typingmind:get-started-with-self-host-installation-instruct, typingmind:typingmind:setup-external-connectors-for-knowledge-base ### DATA-03: Training use and provider retention Judgment rule: Record whether customer content trains vendor or provider models, by plan and default; zero-data-retention arrangements with model providers; abuse-monitoring retention and human review; and the model provider's terms that apply. A "no training" claim on a marketing page is `vendor-claimed` until the contract or DPA says so. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial The terms prohibit model training on customer inputs. API zero retention has a narrow signed secrecy-agreement scope. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client, AlpineAI API platform; version=not specified; plan=Current hosted plans; zero retention only AI API services under signed secret-data supplement; add_ons=Signed supplementary agreement for legally secret data, where applicable; deployment_mode=not specified; region=CH for plaintext secret-data services; general DPA CH/EEA with conditional third countries; os=not specified. Limitations: - Ordinary-chat/provider retention and general staff-review practice outside the signed secret-data scope are not documented. DPA anonymised-data use must be read alongside the narrower no-training clause. - vendor_training — Does the vendor train on customer content? [supported]: AGB 4.4 says customer inputs do not feed back into language-model development. DPA 2.4 separately permits anonymisation and use for service/product development; this is not a blanket prohibition on every secondary use. Sources: swissgpt:swissgpt:terms, swissgpt:swissgpt:dpa. - provider_training — Do model providers train on it under this product's contracts? [partial]: Vendor says models run on AlpineAI hardware without foreign model APIs; a separate third-party model-provider training contract is not established. Sources: swissgpt:swissgpt:product. - provider_retention — How long do model providers keep prompts, including for abuse monitoring? [partial]: Signed secrecy agreement forbids storing/logging AI API input and output for secret data. Ordinary web-chat and general API retention periods are unspecified. Sources: swissgpt:swissgpt:secrecy. - zero_retention — Is zero data retention available, and on which plan? [supported]: Available contractually for AI API services processing legally secret data when the supplementary agreement is signed. It is not a general chat-history promise. Sources: swissgpt:swissgpt:secrecy. - human_review — Can provider staff review content? [supported]: For secret data under the signed supplement, plaintext access needs individual approval, unavoidable BCM emergency or a binding Swiss order; staff abuse monitoring with plaintext access is forbidden. Sources: swissgpt:swissgpt:secrecy. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:terms, swissgpt:swissgpt:dpa, swissgpt:swissgpt:product, swissgpt:swissgpt:secrecy #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial TypingMind asserts no vendor training; customer provider contracts control training and retention. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=TypingMind Teams SaaS; Starter, Growth and Professional prerequisites stated per answer; Business quote not obtained; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - vendor_training — Does the vendor train on customer content? [supported]: The published policy says TypingMind does not use conversations for AI training. This is a vendor policy assertion, not an independently verified contract-wide audit. Sources: typingmind:typingmind:data-usage-and-model-training-policy. - provider_training — Do model providers train on it under this product's contracts? [partial]: Teams uses customer provider keys and direct customer provider contracts. The policy summarizes provider-specific training defaults; this pass did not verify every provider contract or free/paid API variant. Sources: typingmind:typingmind:set-up-api-keys, typingmind:typingmind:data-usage-and-model-training-policy. - provider_retention — How long do model providers keep prompts, including for abuse monitoring? [partial]: The policy discusses provider-dependent abuse monitoring, retention and paid/unpaid distinctions; no uniform current retention duration applies to every customer-selected route. Sources: typingmind:typingmind:data-usage-and-model-training-policy. - zero_retention — Is zero data retention available, and on which plan? [partial]: A customer may select a provider/endpoint with suitable terms, but no TypingMind-wide ZDR entitlement, plan or complete exception matrix was established. Sources: typingmind:typingmind:data-usage-and-model-training-policy, typingmind:typingmind:eu-data-center. - human_review — Can provider staff review content? [partial]: TypingMind’s provider summary says OpenAI authorized employees may review API data for abuse/legal compliance. Its Gemini section distinguishes paid no-training from unpaid improvement use and says human reviewers may read/annotate/process inputs and outputs for quality improvement. These are qualified vendor summaries, not verified terms for every customer paid/unpaid endpoint, ZDR approval or actual account. Sources: typingmind:typingmind:data-usage-and-model-training-policy. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:data-usage-and-model-training-policy, typingmind:typingmind:set-up-api-keys, typingmind:typingmind:eu-data-center ### DATA-04: Retention and deletion Judgment rule: Record default retention for chats, files, indexes and logs; admin-configured retention; user deletion; deletion after contract end; and deletion confirmation. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial The DPA requires deletion after contract end with a backup-cycle exception. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - Default chat/file/index/log periods, customer retention settings, permanent user deletion and the backup-cycle deadline are not specified. - default_retention — What is retained and for how long by default? [unknown]: What is retained and for how long by default? Unknown. Default chat/file/index/log periods, customer retention settings, permanent user deletion and the backup-cycle deadline are not specified. - admin_retention — Can admins set retention periods? [unknown]: Can admins set retention periods? Unknown. Default chat/file/index/log periods, customer retention settings, permanent user deletion and the backup-cycle deadline are not specified. - user_deletion — Can users delete their chats and files permanently? [unknown]: Can users delete their chats and files permanently? Unknown. Default chat/file/index/log periods, customer retention settings, permanent user deletion and the backup-cycle deadline are not specified. - contract_end_deletion — What is deleted at contract end, and when? [partial]: DPA 9.1 requires deletion of provided customer data on termination. Backup copies delete on restoration and ultimately as backup cycles expire; no cycle duration is given. AGB 7.5 offers written confirmation on request with legal-retention exceptions. Sources: swissgpt:swissgpt:dpa, swissgpt:swissgpt:terms. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:dpa, swissgpt:swissgpt:terms #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Admins control recorded chat retention; termination deletion/return is contractual without a fixed public completion window. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams Professional for admin chat recording/retention; general service privacy/DPA rights separate from personal/local history; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - default_retention — What is retained and for how long by default? [partial]: Professional admin chat recording is off by default; that is separate from user-visible history and personal browser storage. The privacy policy separately mentions90-day unpaid-account inactivity deletion with warning; a paid Teams history lifetime is not specified. Sources: typingmind:typingmind:export-chat-logs, typingmind:typingmind:privacy-policy, typingmind:typingmind:r2-plan-matrix. - admin_retention — Can admins set retention periods? [supported]: Professional chat-log settings set recorded-chat retention, with permanent deletion of older records. No personal/local or backup deletion interval is inferred. Sources: typingmind:typingmind:view-and-export-user-chat-logs, typingmind:typingmind:r2-plan-matrix. - user_deletion — Can users delete their chats and files permanently? [partial]: The privacy policy provides data-erasure rights; immediate permanent deletion across Teams files, backups and logs by the user UI was not established. Sources: typingmind:typingmind:privacy-policy. - contract_end_deletion — What is deleted at contract end, and when? [partial]: The DPA permits customer-choice deletion or return and certification on request, subject to legal retention; a fixed completion deadline and self-serve export window are not specified. Sources: typingmind:typingmind:data-processing-agreement. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:export-chat-logs, typingmind:typingmind:privacy-policy, typingmind:typingmind:r2-plan-matrix, typingmind:typingmind:view-and-export-user-chat-logs, typingmind:typingmind:data-processing-agreement ### DATA-05: Sub-processors and foreign-law exposure Judgment rule: Record the published sub-processor list with countries, change-notification rules, the vendor's and parent company's jurisdiction, transfer mechanisms (EU-US Data Privacy Framework, Swiss-US DPF, SCCs) and the vendor's documented position on foreign government access requests (e.g. US CLOUD Act). Record facts, not a legal assessment. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial A public July 2026 DPA annex lists operational suppliers and contract notification/transfer terms. It is a dated source, not proof of the complete current provider list: RZO appears only on the security page and its current contractual status remains unknown. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client, AlpineAI API platform; version=DPA Annex 3 dated 10 July 2026; re-retrieved 21 October 2026; plan=AlpineAI platform contracts; current tenant-specific supplier schedule not inspected; add_ons=none stated; deployment_mode=not specified; region=Dated annex CH/EU/US supplier locations; current complete country/supplier schedule unverified; os=not specified. Limitations: - No current complete provider/country schedule or RZO-specific authorisation was retrieved; no inference is made that RZO is approved, unauthorised or a separate model provider. - Ultimate ownership and supplier-specific US transfer instruments remain unknown. - subprocessor_list — Is a dated sub-processor list published? [partial]: A public Annex 3 dated 10 July 2026 is downloadable through /avv/. It includes RIZ hosting/support and Green colocation. The security page additionally names RZO. No updated public annex, RZO approval notice or tenant-specific schedule was obtained, so completeness/current applicability is unknown. Sources: swissgpt:swissgpt:review2-dpa, swissgpt:swissgpt:review2-security. - subprocessor_countries — Which countries do sub-processors operate in? [partial]: The dated annex lists CH data locations for RIZ, Green, cloudscale, accounting and IT operations; EU for helpdesk/payments/customer success; US for CRM and opt-in Brave search. Green is described as rack/power/network colocation without logical access, not model-provider API processing. These are the annex entries, not an exhaustive current content-processing country map. Sources: swissgpt:swissgpt:review2-dpa. - change_notification — Are customers notified of sub-processor changes? [supported]: DPA 7.3 requires text notice of intended additions/replacements and gives 14 days for an objection supported by an important reason. No customer-specific change notice was inspected. Sources: swissgpt:swissgpt:review2-dpa. - parent_jurisdiction — Which jurisdiction governs the vendor and its parent? [partial]: The contracting entity is AlpineAI AG in Davos, Switzerland; ultimate parent/beneficial-owner jurisdiction was not established. Sources: swissgpt:swissgpt:imprint, swissgpt:swissgpt:terms. - transfer_mechanism — Which transfer mechanism applies to non-EU/CH processing? [partial]: DPA 1.5 conditions third-country processing on prior written notice and adequate safeguards such as SCCs or adequacy. It does not map a particular transfer instrument to each US supplier. Sources: swissgpt:swissgpt:review2-dpa. - government_access_policy — Is there a published government-access or transparency policy? [supported]: DPA 2.1 requires notice of legally mandated processing unless prohibited. The separately signed secret-data supplement adds Swiss-order notice, challenge assistance and minimal disclosure, within its Swiss secret-data scope. Sources: swissgpt:swissgpt:review2-dpa, swissgpt:swissgpt:secrecy. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:review2-dpa, swissgpt:swissgpt:review2-security, swissgpt:swissgpt:imprint, swissgpt:swissgpt:terms, swissgpt:swissgpt:secrecy #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial A public DPA lists processors and transfer measures; vendor legal jurisdiction differs from hosting region. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=TypingMind Teams SaaS; Starter, Growth and Professional prerequisites stated per answer; Business quote not obtained; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - subprocessor_list — Is a dated sub-processor list published? [partial]: DPA Appendix II publishes a subprocessor list; no standalone dated list revision was visible in the retrieved document. Sources: typingmind:typingmind:data-processing-agreement. - subprocessor_countries — Which countries do sub-processors operate in? [supported]: The list includes AWS US/Germany, Render US, Vercel US/EU, US services including Sentry/Axiom/Customer.io/Stripe/HelpScout/Mailgun and Paddle US/UK. EU storage does not remove these separate vendor paths. Sources: typingmind:typingmind:data-processing-agreement. - change_notification — Are customers notified of sub-processor changes? [supported]: DPA 7.8 says the processor will endeavor to give 30 days’ notice of additions/replacements, with 15 days for objections; preserve the qualified notice wording. Sources: typingmind:typingmind:data-processing-agreement. - parent_jurisdiction — Which jurisdiction governs the vendor and its parent? [partial]: Terms identify Typing Technology Co Ltd with Singapore governing law; the DPA gives a Vietnam processor address. Parent ownership/jurisdiction and actual staff locations were not independently established. Sources: typingmind:typingmind:terms-of-service, typingmind:typingmind:data-processing-agreement. - transfer_mechanism — Which transfer mechanism applies to non-EU/CH processing? [supported]: DPA 7.9 and its SCC appendix provide EU transfer safeguards, subject to actual processing configuration; a Swiss-specific addendum was not established. Sources: typingmind:typingmind:data-processing-agreement. - government_access_policy — Is there a published government-access or transparency policy? [partial]: The privacy policy allows disclosures for authorized legal requests; no separate transparency report or detailed government-request policy was retrieved. Sources: typingmind:typingmind:privacy-policy. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:data-processing-agreement, typingmind:typingmind:terms-of-service, typingmind:typingmind:privacy-policy ### DATA-06: Contracts and regulatory coverage Judgment rule: Record DPA availability and form (click-through, negotiable), reference to Swiss revDSG and GDPR, governing law and court, EU AI Act documentation (provider or deployer role, transparency materials), and sector-specific materials (FINMA outsourcing, banking secrecy, public sector). No compliance verdict. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial Public DPA and secrecy supplement address Swiss data protection and professional/official secrecy. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - DPA negotiability, EU AI Act provider/deployer documents and a FINMA-specific pack are not provided in the reviewed public materials. - dpa — Is a DPA available, and is it negotiable? [partial]: A public DPA applies to individuals without AlpineAI signature; companies may obtain a countersigned copy. Negotiation rights are not established. Sources: swissgpt:swissgpt:dpa. - revdsg_gdpr — Does the DPA cover revDSG and GDPR? [supported]: DPA expressly cites Art. 9 Swiss DSG and Art. 28 GDPR and defines controller/processor roles. Sources: swissgpt:swissgpt:dpa. - governing_law — Which law and court govern the contract? [supported]: DPA 11.4 specifies Swiss law and exclusive Zürich jurisdiction; AGB 11.4 preserves AlpineAI rights to sue elsewhere. Sources: swissgpt:swissgpt:dpa, swissgpt:swissgpt:terms. - ai_act_materials — Is EU AI Act documentation provided? [unknown]: Is EU AI Act documentation provided? Unknown. DPA negotiability, EU AI Act provider/deployer documents and a FINMA-specific pack are not provided in the reviewed public materials. - sector_materials — Are FINMA or other sector materials provided? [partial]: Public signed supplement covers legally secret data including official/professional secrecy, CH plaintext services and annual assurance evidence. No FINMA-specific outsourcing dossier was located. Sources: swissgpt:swissgpt:secrecy. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:dpa, swissgpt:swissgpt:terms, swissgpt:swissgpt:secrecy #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial A public GDPR DPA and Singapore-law business terms are available; Swiss and sector-specific undertakings remain gaps. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=TypingMind Teams SaaS; Starter, Growth and Professional prerequisites stated per answer; Business quote not obtained; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - dpa — Is a DPA available, and is it negotiable? [partial]: A public fillable GDPR Article 28 DPA names the processor, security measures and SCCs. Whether bespoke negotiation is available was not established. Sources: typingmind:typingmind:data-processing-agreement. - revdsg_gdpr — Does the DPA cover revDSG and GDPR? [partial]: The DPA expressly addresses GDPR and SCCs; explicit revised Swiss FADP/revDSG coverage was not established. Sources: typingmind:typingmind:data-processing-agreement. - governing_law — Which law and court govern the contract? [supported]: The September 17, 2026 Terms use Singapore law, with SIAC English-language business arbitration; the DPA SCC appendix separately uses EU exporter-state law/forum where applicable. Sources: typingmind:typingmind:terms-of-service, typingmind:typingmind:data-processing-agreement. - ai_act_materials — Is EU AI Act documentation provided? [unknown]: Is EU AI Act documentation provided? No EU AI Act compliance package, FINMA outsourcing materials or dedicated professional-secrecy agreement was retrieved. - sector_materials — Are FINMA or other sector materials provided? [unknown]: Are FINMA or other sector materials provided? No EU AI Act compliance package, FINMA outsourcing materials or dedicated professional-secrecy agreement was retrieved. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:data-processing-agreement, typingmind:typingmind:terms-of-service ### DATA-07: Certifications and attestations Judgment rule: Record each certificate or report separately: ISO 27001, 27017, 27018, 27701, 42001; SOC 2 Type I/II with period; BSI C5; TISAX; national schemes. Record issuer, scope (which product and entity), validity and access route. A cloud provider's certificate does not certify the product. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial AlpineAI claims ISO 27001:2022 and VDSZ issued by SQS; certificates were not inspected. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - No issuer certificate or full report was obtained. ISO/VDSZ validity and scope, SOC 2 type/period, C5 and independent penetration-test summary remain unverified. - iso_27001 — ISO 27001 with scope and validity? [partial]: Vendor identifies AlpineAI AG, SQS and ISO 27001:2022. Certificate scope and validity were not retrieved, so this remains vendor-claimed. Sources: swissgpt:swissgpt:security. - soc2 — SOC 2 type, period and covered service? [unknown]: SOC 2 type, period and covered service? Unknown. No issuer certificate or full report was obtained. ISO/VDSZ validity and scope, SOC 2 type/period, C5 and independent penetration-test summary remain unverified. - c5 — BSI C5 attestation? [unknown]: BSI C5 attestation? Unknown. No issuer certificate or full report was obtained. ISO/VDSZ validity and scope, SOC 2 type/period, C5 and independent penetration-test summary remain unverified. - other_certifications — Which other certificates (27701, 42001, TISAX)? [partial]: Vendor claims VDSZ, issued by SQS. No certificate number, scope or expiry was retrieved. Sources: swissgpt:swissgpt:security. - report_access — How can customers obtain reports? [supported]: DPA 6.1 permits documentary assurance such as ISO records/internal audits; signed secret-data supplement promises annual certification/report covering the processing without extra charge. Sources: swissgpt:swissgpt:dpa, swissgpt:swissgpt:secrecy. - pentest — Are independent penetration test summaries available? [partial]: DPA Annex 2 promises regular penetration testing; independence and a public summary are not established. Sources: swissgpt:swissgpt:dpa. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:security, swissgpt:swissgpt:dpa, swissgpt:swissgpt:secrecy #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial The self-host page claims SOC 2 Type II. A report with auditor, period and covered-service scope was not retrieved, and other attestations remain unverified. Availability: generally-available Delivery: native Scope: surfaces=Commercial self-hosted TypingMind Teams server; version=Public self-host page retrieved 2026-10-26; no certificate verified; plan=Commercial Teams self-host marketing assurance claim; no customer report or SaaS certificate scope inferred; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=not specified; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Vendor assertion only. No independently verified report, certificate scope or validity was accessed; trust.typingmind.com returned HTTP 403. - iso_27001 — ISO 27001 with scope and validity? [unknown]: ISO 27001 with scope and validity? Trust Center access returned HTTP 403. The retrieved DPA security appendix is not an ISO/SOC/C5 certificate. Issuer, service scope, validity/report period, report-access process and independent penetration-test summary were not retrieved. - soc2 — SOC 2 type, period and covered service? [partial]: The vendor advertises SOC 2 Type II on the Teams self-host page. Auditor, report period, covered entity/service and applicability to a customer-operated deployment are unverified. Sources: typingmind:typingmind:r4-soc2-selfhost-claim. - c5 — BSI C5 attestation? [unknown]: BSI C5 attestation? Trust Center access returned HTTP 403. The retrieved DPA security appendix is not an ISO/SOC/C5 certificate. Issuer, service scope, validity/report period, report-access process and independent penetration-test summary were not retrieved. - other_certifications — Which other certificates (27701, 42001, TISAX)? [unknown]: Which other certificates (27701, 42001, TISAX)? Trust Center access returned HTTP 403. The retrieved DPA security appendix is not an ISO/SOC/C5 certificate. Issuer, service scope, validity/report period, report-access process and independent penetration-test summary were not retrieved. - report_access — How can customers obtain reports? [unknown]: How can customers obtain reports? Trust Center access returned HTTP 403. The retrieved DPA security appendix is not an ISO/SOC/C5 certificate. Issuer, service scope, validity/report period, report-access process and independent penetration-test summary were not retrieved. - pentest — Are independent penetration test summaries available? [unknown]: Are independent penetration test summaries available? Trust Center access returned HTTP 403. The retrieved DPA security appendix is not an ISO/SOC/C5 certificate. Issuer, service scope, validity/report period, report-access process and independent penetration-test summary were not retrieved. Checked 2026-10-26; agent cross-checked 2026-10-26; check again by 2026-12-25. Sources: typingmind:typingmind:r4-soc2-selfhost-claim ### DATA-08: Encryption and key control Judgment rule: Record encryption at rest and in transit, customer-managed keys (BYOK/HYOK), key location, and private networking (private link, IP allow-lists, VPN). #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial The vendor describes AES-256 at rest, TLS in transit and isolation on its own hardware. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - Customer-managed or customer-held keys, key location, customer private connectivity and technical tenant-isolation model remain unspecified. - encryption_at_rest — Is data encrypted at rest? [supported]: Security page names AES-256 at rest and TLS between clients and platform; DPA Annex 2 requires encryption. Sources: swissgpt:swissgpt:security, swissgpt:swissgpt:dpa. - customer_managed_keys — Can the customer manage or hold the keys? [unknown]: Can the customer manage or hold the keys? Unknown. Customer-managed or customer-held keys, key location, customer private connectivity and technical tenant-isolation model remain unspecified. - private_networking — Is private connectivity or IP restriction available? [partial]: GPU partners are described as VPN-only infrastructure. A customer Private Link, VPN or IP allow-list option is not established. Sources: swissgpt:swissgpt:security. - tenant_isolation — How are customers separated in shared SaaS? [partial]: Vendor claims isolated/private-cloud environments and least-privilege RBAC. Exact shared-SaaS tenant separation and single-tenancy guarantees are not documented. Sources: swissgpt:swissgpt:security. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:security, swissgpt:swissgpt:dpa #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Encryption and private self-host infrastructure are documented without a SaaS customer-key guarantee. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel, Commercial self-hosted TypingMind Teams server, TypingMind personal browser app and PWA; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams SaaS US/EU storage; separately quoted annual Teams self-host network controls; personal local-key claims explicitly separate; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=not specified; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Teams OpenAI route: server requests using centrally configured admin keys. Personal direct-browser/local-key assertions are not inherited by Teams. - The generic policy versus Teams FAQ scope discrepancy remains visible pending policy clarification. - encryption_at_rest — Is data encrypted at rest? [partial]: The EU guide says stored information is encrypted; exact cipher, key ownership and component coverage were not established. Personal optional cloud encryption is a different surface. Sources: typingmind:typingmind:eu-data-center, typingmind:typingmind:privacy-policy. - customer_managed_keys — Can the customer manage or hold the keys? [unknown]: Can the customer manage or hold the keys? No Teams SaaS customer-managed/held encryption-key option or complete private-connectivity configuration was established. - private_networking — Is private connectivity or IP restriction available? [partial]: Self-host permits customer network placement; localhost/private custom endpoints require ALLOW_UNSAFE_USER_ENDPOINTS. SaaS private links, IP allowlisting and private model routing were not established. Sources: typingmind:typingmind:get-started-with-self-host-installation-instruct. - tenant_isolation — How are customers separated in shared SaaS? [partial]: The privacy policy calls Teams multitenant; logical isolation implementation and independent tenant-boundary testing were not established. Sources: typingmind:typingmind:privacy-policy. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:eu-data-center, typingmind:typingmind:privacy-policy, typingmind:typingmind:get-started-with-self-host-installation-instruct, typingmind:typingmind:r2-teams-server-route ### DATA-09: Professional secrecy Judgment rule: For professional secrecy holders (Berufsgeheimnisträger: lawyers, doctors, auditors, tax advisers) under § 203 StGB (DE) and Art. 321 StGB (CH), record whether the vendor signs a contractual secrecy undertaking or a § 203 agreement that binds its staff and its sub-processors; whether this comes as a dedicated offering or a paid add-on, and on which plan; who can access content in plain text (vendor staff, sub-processors, model providers); and the documentation provided. A DPA or a no-training clause alone is not a secrecy undertaking. Banking secrecy and FINMA outsourcing belong to DATA-06, not here. Record contract facts, not a legal assessment. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial AlpineAI publishes a separate bilateral undertaking for Swiss official, professional and other statutory secret data. It binds staff and subcontractors, restricts plaintext access and requires Swiss plaintext services when signed. Plan eligibility, any supplement fee and German § 203 coverage remain unverified. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client, AlpineAI API platform; version=Supplement printed 30 March 2026; plan=AlpineAI platform under a signed main contract and supplementary secrecy agreement; eligible plan tiers unpublished; add_ons=Signed supplementary secrecy agreement; price unpublished; deployment_mode=not specified; region=Swiss statutory secrecy; services exposing secret data in plaintext, including APIs, exclusively in Switzerland; os=not specified. Limitations: - The supplement is conditional on both signatures and processing Swiss legally secret data. Its Swiss Art. 320 ff. reference is not evidence of a German § 203 agreement or a legal suitability assessment for a particular profession. - Plan eligibility, any supplement fee, executed customer agreements and the actual downstream staff/subcontractor agreements remain unverified. - secrecy_undertaking — Does the vendor sign a secrecy undertaking or § 203 StGB / Art. 321 StGB agreement covering its own staff? [supported]: The public two-party signature template supplements the platform main contract and DPA. It takes effect on the last signature and only while the customer processes data subject to Swiss official, professional or other statutory secrecy (example: Art. 320 ff. Swiss StGB). Section 3 requires equivalent contractual obligations for AlpineAI staff who may access plaintext. This establishes a published undertaking for signature, not an inspected executed customer agreement or an express German § 203 instrument. Sources: swissgpt:swissgpt:secrecy. - subprocessor_coverage — Does the undertaking extend to its sub-processors, and is that documented? [supported]: Section 3 requires AlpineAI to impose at least equivalent contractual obligations on its subcontractors and staff who may access secret data in plaintext. The DPA safeguards also apply to secret data, with the stronger secrecy supplement taking priority. Actual downstream signed agreements were not public and were not inspected. Sources: swissgpt:swissgpt:secrecy. - secrecy_offering — Is it a dedicated offering or add-on, and on which plan? [partial]: The security page offers corporate customers an additional secrecy agreement; the public template is separate from the main contract and DPA and requires both parties to sign. No named eligible package, paid add-on tariff or general supplement fee was established. The annual assurance evidence is promised without extra cost, which does not establish that the agreement itself is free. Sources: swissgpt:swissgpt:security, swissgpt:swissgpt:secrecy. - content_access — Who can access content in plain text: vendor staff, sub-processors, model providers? [supported]: Under the signed supplement, AlpineAI and subcontractor staff may access plaintext only on a need-to-know basis plus individual prior customer approval, an unavoidable urgent business-continuity emergency, or a binding Swiss authority order. AlpineAI must take reasonable measures to prevent third-party plaintext access. AI API inputs and outputs must not be stored or logged, human abuse monitoring with plaintext access is prohibited, and plaintext services including APIs must remain in Switzerland. The security page separately says AlpineAI operates the models and colocation providers have no logical access to systems or data. These are contract terms and vendor disclosures, not a hands-on verification of every data flow. Sources: swissgpt:swissgpt:secrecy, swissgpt:swissgpt:security. - secrecy_documentation — Is the offering publicly documented (contract template, data-flow description, professional-body guidance)? [supported]: A public two-page contract template, printed 30 March 2026, sets scope, precedence, secrecy duties, staff/subcontractor flow-down, plaintext access, Swiss processing, API retention and signature blocks. Section 3 also promises annual assurance evidence covering secret-data processing at no extra cost and a customer audit right at customer expense. The security page explains the supplementary agreement. No executed customer contract was inspected. Sources: swissgpt:swissgpt:secrecy, swissgpt:swissgpt:security. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:secrecy, swissgpt:swissgpt:security #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial General GDPR confidentiality and configurable admin/model access are documented; a specific §203 or Art321 undertaking is unverified. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel, Commercial self-hosted TypingMind Teams server; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams SaaS Professional chat governance and separately quoted self-host licence; ordinary public DPA; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=not specified; region=US/EU SaaS or customer self-host placement; customer provider contracts separate; os=not specified. Limitations: - DPA confidentiality, direct provider contracts and self-host placement alone do not establish professional-secrecy suitability. - No legal suitability or compliance conclusion. - secrecy_undertaking — Does the vendor sign a secrecy undertaking or § 203 StGB / Art. 321 StGB agreement covering its own staff? [unknown]: DPA7.4(c) limits personnel access and requires confidentiality/statutory duties. No signed/signon §203 StGB or Art321 StGB professional-secrecy undertaking was retrieved; ordinary GDPR confidentiality is not treated as that agreement. - subprocessor_coverage — Does the undertaking extend to its sub-processors, and is that documented? [unknown]: DPA7.8(d)-(e) requires substantially equivalent data-protection obligations and allows a redacted subprocessor agreement on request. A dedicated professional-secrecy undertaking flowing to subprocessors was not established. - secrecy_offering — Is it a dedicated offering or add-on, and on which plan? [unknown]: No specific professional-secrecy package/add-on or SaaS/self-host plan entitlement was established. Customer-operated infrastructure and the commercial self-host licence do not by themselves prove that undertaking. - content_access — Who can access content in plain text: vendor staff, sub-processors, model providers? [partial]: SaaS authorized personnel may access data only as needed under the DPA; customer admins with chat-log permission may inspect recorded prompts/answers when logging is enabled. Selected customer API providers process content; self-host placement changes operator access but model/embedding/blob routes remain relevant. A complete plaintext support/subprocessor/provider human-access matrix was not retrieved. Sources: typingmind:typingmind:dpa-confidentiality, typingmind:typingmind:view-and-export-user-chat-logs, typingmind:typingmind:roles-and-permissions, typingmind:typingmind:set-up-api-keys, typingmind:typingmind:get-started-with-self-host-installation-instruct. - secrecy_documentation — Is the offering publicly documented (contract template, data-flow description, professional-body guidance)? [unknown]: Public DPA, security appendix, processor list and self-host data-flow guides were retrieved. Targeted first-party search did not establish a professional-secrecy contract template, DE/CH-specific data-flow addendum or professional-body guidance. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:dpa-confidentiality, typingmind:typingmind:view-and-export-user-chat-logs, typingmind:typingmind:roles-and-permissions, typingmind:typingmind:set-up-api-keys, typingmind:typingmind:get-started-with-self-host-installation-instruct ## Commercials and vendor ### VEN-01: Pricing model and list prices Judgment rule: Record pricing unit (per seat, per active user, usage-based, platform fee, hybrid), list prices per plan in original currency with date, included usage and credits, add-on prices (connectors, agents, image generation) and overage rates. Cite public prices or dated quotes. Per-seat versus usage is a fact, not a quality judgment. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial Prices differ by sector and billing period; Enterprise is quoted from 10 licences. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - Enterprise rates, VAT treatment, university monthly-toggle verification, model/search quotas, overage prices and integration charges are not established. Do not treat the annual default shown by text extraction as monthly billing. - pricing_unit — What is the pricing unit? [supported]: Standard/Premium price per individual user licence. Enterprise package starts at 10 licences. Sources: swissgpt:swissgpt:public-pricing, swissgpt:swissgpt:health-pricing, swissgpt:swissgpt:university. - list_prices — What are list prices per plan, with date and currency? [partial]: As retrieved 2026-10-21: administration Standard/Premium CHF 79/89 per user/month with monthly billing, CHF 74/80 per month with annual billing. Healthcare CHF 89/99 monthly or CHF 80/90 per month annual. University page displays annual CHF 74/80; monthly amounts were not independently toggled. Tax inclusion is not stated clearly, so no ex-VAT conversion is asserted. Sources: swissgpt:swissgpt:public-pricing, swissgpt:swissgpt:health-pricing, swissgpt:swissgpt:university. - included_usage — What usage is included? [partial]: Standard has chat/search, uploads, Outlook and 2 assistants; Premium adds Listener, prebuilt assistants and 5 assistants, each listed at 50 documents / max. 100 MB. General inference/search token allowances are not published. Sources: swissgpt:swissgpt:university. - add_ons — Which features cost extra? [partial]: Premium adds Listener and preconfigured assistants; Enterprise covers custom integrations. AGB 2.3 permits charges for new major functionality; numeric integration/service fees are unspecified. Sources: swissgpt:swissgpt:university, swissgpt:swissgpt:terms. - overage — What do usage beyond limits and premium models cost? [unknown]: What do usage beyond limits and premium models cost? Unknown. Enterprise rates, VAT treatment, university monthly-toggle verification, model/search quotas, overage prices and integration charges are not established. Do not treat the annual default shown by text extraction as monthly billing. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:public-pricing, swissgpt:swissgpt:health-pricing, swissgpt:swissgpt:university, swissgpt:swissgpt:terms #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Teams subscriptions, personal one-time licences and self-host annual quotes are separate cost structures. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel, Commercial self-hosted TypingMind Teams server, TypingMind personal browser app and PWA; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams SaaS Starter/Growth/Professional/quoted Business; separate annual Teams self-host quote; personal Standard/Extended/Premium one-time licences and separate bulk pack; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=not specified; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - pricing_unit — What is the pricing unit? [supported]: Teams SaaS uses recurring instance/seats billing; commercial Teams self-host is an annual quoted licence per instance. Personal one-time licences, including the separate shareable bulk pack, do not establish a Teams administration entitlement. Sources: typingmind:typingmind:pricing, typingmind:typingmind:cloud-host-vs-self-host, typingmind:typingmind:r2-personal-bulk-licence. - list_prices — What are list prices per plan, with date and currency? [supported]: As retrieved October 21, annual-selected landing cards display USD83/166/249 per month Starter/Growth/Professional (Professional crossed-out USD499). The documentation’s monthly table lists USD99/199/299 plus USD8/additional user/month. Business starts USD18,700/year by quote; conditional “as low as USD4/user/month” is not a universal rate. Sources: typingmind:typingmind:pricing, typingmind:typingmind:compare-typingmind-team-plans. - included_usage — What usage is included? [supported]: All SaaS tiers include five users and unlimited configured models/agents/plugins/prompts, excluding provider API usage. One million KB training characters belongs to the Growth/Professional RAG guide, still labeled Early Access; no Starter KB allowance is inferred. Sources: typingmind:typingmind:r2-plan-matrix, typingmind:typingmind:how-knowledge-base-works-in-typingmind-team. - add_ons — Which features cost extra? [supported]: Additional seats, KB character capacity, model APIs and third-party plugin services are separate; self-host SLA/dedicated support is an add-on. Sources: typingmind:typingmind:pricing, typingmind:typingmind:selfhost, typingmind:typingmind:deep-research. - overage — What do usage beyond limits and premium models cost? [partial]: Model usage is billed by the customer’s chosen provider. No unified TypingMind premium-model tariff applies; extra KB capacity and Business/self-host scope require current checkout or quote. Sources: typingmind:typingmind:pricing, typingmind:typingmind:set-up-api-keys. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:pricing, typingmind:typingmind:cloud-host-vs-self-host, typingmind:typingmind:r2-personal-bulk-licence, typingmind:typingmind:compare-typingmind-team-plans, typingmind:typingmind:r2-plan-matrix, typingmind:typingmind:how-knowledge-base-works-in-typingmind-team, typingmind:typingmind:selfhost, typingmind:typingmind:deep-research, typingmind:typingmind:set-up-api-keys ### VEN-02: Scenario cost Judgment rule: Price three workloads for 12 months: 200, 1,000 and 5,000 users with a disclosed light/regular/power usage mix, in CHF and EUR excluding VAT. Include licences, platform fees, add-ons, model usage and, for self-hosted options, a stated operating assumption. Separate fixed, variable, included and unknown costs. No invented discounts. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial ALM-14 shared cost assumptions v1.4, 2026-10-21. The required 200, 1,000 and 5,000-seat annual scenarios cannot be priced from the public individual plans. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - ALM-14 shared cost assumptions v1.4, 2026-10-21. Horizon 12 months, list prices excluding VAT, no invented discounts. Mix 60% light at 100 messages/user/month, 30% regular at 400, 10% power at 1,500. Where model usage is separately billed, each message has 3,000 input and 600 output tokens, giving 990k input plus 198k output per user/month. At 200/1,000/5,000 users this is 198m/990m/4.95bn input and 39.6m/198m/990m output per month. All three seat counts require an unpublished enterprise/custom price; no quote was requested. CHF and EUR ex-VAT totals, fixed charges, variable charges and add-ons remain unknown. FX dated 2026-10-20: EUR 1 = CHF 0.9237 and USD 1.1215; CHF to EUR divide by 0.9237, USD to EUR divide by 1.1215, source ECB. No numeric conversion or proxy model tariff is applied because there is no applicable public enterprise base/usage rate. - cost_200 — Year-one cost at 200 users? [unknown]: ALM-14 shared cost assumptions v1.4, 2026-10-21. Horizon 12 months, list prices excluding VAT, no invented discounts. Mix 60% light at 100 messages/user/month, 30% regular at 400, 10% power at 1,500. Where model usage is separately billed, each message has 3,000 input and 600 output tokens, giving 990k input plus 198k output per user/month. At 200/1,000/5,000 users this is 198m/990m/4.95bn input and 39.6m/198m/990m output per month. All three seat counts require an unpublished enterprise/custom price; no quote was requested. CHF and EUR ex-VAT totals, fixed charges, variable charges and add-ons remain unknown. FX dated 2026-10-20: EUR 1 = CHF 0.9237 and USD 1.1215; CHF to EUR divide by 0.9237, USD to EUR divide by 1.1215, source ECB. No numeric conversion or proxy model tariff is applied because there is no applicable public enterprise base/usage rate. - cost_1000 — Year-one cost at 1,000 users? [unknown]: ALM-14 shared cost assumptions v1.4, 2026-10-21. Horizon 12 months, list prices excluding VAT, no invented discounts. Mix 60% light at 100 messages/user/month, 30% regular at 400, 10% power at 1,500. Where model usage is separately billed, each message has 3,000 input and 600 output tokens, giving 990k input plus 198k output per user/month. At 200/1,000/5,000 users this is 198m/990m/4.95bn input and 39.6m/198m/990m output per month. All three seat counts require an unpublished enterprise/custom price; no quote was requested. CHF and EUR ex-VAT totals, fixed charges, variable charges and add-ons remain unknown. FX dated 2026-10-20: EUR 1 = CHF 0.9237 and USD 1.1215; CHF to EUR divide by 0.9237, USD to EUR divide by 1.1215, source ECB. No numeric conversion or proxy model tariff is applied because there is no applicable public enterprise base/usage rate. - cost_5000 — Year-one cost at 5,000 users? [unknown]: ALM-14 shared cost assumptions v1.4, 2026-10-21. Horizon 12 months, list prices excluding VAT, no invented discounts. Mix 60% light at 100 messages/user/month, 30% regular at 400, 10% power at 1,500. Where model usage is separately billed, each message has 3,000 input and 600 output tokens, giving 990k input plus 198k output per user/month. At 200/1,000/5,000 users this is 198m/990m/4.95bn input and 39.6m/198m/990m output per month. All three seat counts require an unpublished enterprise/custom price; no quote was requested. CHF and EUR ex-VAT totals, fixed charges, variable charges and add-ons remain unknown. FX dated 2026-10-20: EUR 1 = CHF 0.9237 and USD 1.1215; CHF to EUR divide by 0.9237, USD to EUR divide by 1.1215, source ECB. No numeric conversion or proxy model tariff is applied because there is no applicable public enterprise base/usage rate. - variable_share — Which part depends on usage? [unknown]: ALM-14 shared cost assumptions v1.4, 2026-10-21. Horizon 12 months, list prices excluding VAT, no invented discounts. Mix 60% light at 100 messages/user/month, 30% regular at 400, 10% power at 1,500. Where model usage is separately billed, each message has 3,000 input and 600 output tokens, giving 990k input plus 198k output per user/month. At 200/1,000/5,000 users this is 198m/990m/4.95bn input and 39.6m/198m/990m output per month. All three seat counts require an unpublished enterprise/custom price; no quote was requested. CHF and EUR ex-VAT totals, fixed charges, variable charges and add-ons remain unknown. FX dated 2026-10-20: EUR 1 = CHF 0.9237 and USD 1.1215; CHF to EUR divide by 0.9237, USD to EUR divide by 1.1215, source ECB. No numeric conversion or proxy model tariff is applied because there is no applicable public enterprise base/usage rate. - unknown_parts — Which costs could not be priced? [supported]: Enterprise applies from 10 licences and is quote-only; fixed package, usage, integrations, service charges and tax treatment remain unknown. No sales quote was requested. Sources: swissgpt:swissgpt:university. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:university, swissgpt:swissgpt:ecb-fx #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial A conditional Professional monthly licence-plus-BYOK floor is calculable; Business and self-host totals remain unknown. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Conditional documented Professional monthly list route; not a large-scale quote; add_ons=Model usage proxy; other add-ons unknown; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Known price floor plus explicitly unpriced parts; no complete year-one TCO or negotiated price claim. - cost_200 — Year-one cost at 200 users? [partial]: ALM-14 shared cost assumptions v1.4, 2026-10-21: 12 months ex VAT, 60% light100/30% regular400/10% power1500 messages per user/month, hence330; 3000 input+600 output tokens/message. Conditional undiscounted Teams Professional monthly route USD299 + USD8*(N−5) monthly, not the annual promotion or a Business quote. Direct-provider BYOK proxy gpt-6-luna, reasoning effort none, Standard short context USD0.10 input/USD0.50 output per million; no cache => USD2.376/user/year. ECB2026-10-20 EUR1=USD1.1215=CHF0.9237; USD/EUR divide1.1215, EUR/CHF multiply0.9237. For200 users: licence USD22,308 + model proxy USD475.20 = from USD22,783.20 / EUR20,314.94 / CHF18,764.91 + unknown components. Not a 200-seat quote. Sources: typingmind:typingmind:compare-typingmind-team-plans, typingmind:typingmind:pricing, typingmind:typingmind:openai-price, typingmind:typingmind:openai-model, typingmind:typingmind:fx. - cost_1000 — Year-one cost at 1,000 users? [partial]: Same assumptions:1000 users licence USD99,108 + model USD2,376 = from USD101,484 / EUR90,489.52 / CHF83,585.17 + unknown components. Scale entitlement and negotiated Business alternative unverified. Sources: typingmind:typingmind:compare-typingmind-team-plans, typingmind:typingmind:pricing, typingmind:typingmind:openai-price, typingmind:typingmind:openai-model, typingmind:typingmind:fx. - cost_5000 — Year-one cost at 5,000 users? [partial]: Same assumptions:5000 users licence USD483,108 + model USD11,880 = from USD494,988 / EUR441,362.46 / CHF407,686.51 + unknown components. Scale entitlement and Business alternative unverified. Sources: typingmind:typingmind:compare-typingmind-team-plans, typingmind:typingmind:pricing, typingmind:typingmind:openai-price, typingmind:typingmind:openai-model, typingmind:typingmind:fx. - variable_share — Which part depends on usage? [supported]: The disclosed model proxy contributes USD2.376/user/year. Actual provider model mix, tools, images, audio, embeddings, cached/context tokens and regional API uplifts change usage. This proxy is a comparison assumption, not a vendor default or invoice. Sources: typingmind:typingmind:set-up-api-keys, typingmind:typingmind:openai-price, typingmind:typingmind:openai-model. - unknown_parts — Which costs could not be priced? [supported]: Unknown: large-scale SaaS entitlement/Business quote, current checkout promotion terms, extra KB capacity, connectors/search/plugin bills, embeddings, support/implementation and optional regional processing uplift. Self-host annual commercial licence/support quote plus MySQL, optional PostgreSQL/blob/network infrastructure and operator effort are unpriced; no zero-licence OSS assumption applies. Sources: typingmind:typingmind:pricing, typingmind:typingmind:selfhost, typingmind:typingmind:get-started-with-self-host-installation-instruct, typingmind:typingmind:deep-research. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:compare-typingmind-team-plans, typingmind:typingmind:pricing, typingmind:typingmind:openai-price, typingmind:typingmind:openai-model, typingmind:typingmind:fx, typingmind:typingmind:set-up-api-keys, typingmind:typingmind:selfhost, typingmind:typingmind:get-started-with-self-host-installation-instruct, typingmind:typingmind:deep-research ### VEN-03: Evaluation access Judgment rule: Distinguish free tier, self-serve trial, sales-led pilot, paid proof of concept and open-source self-install. Record card requirement, duration, seat limit, and which features (SSO, connectors, admin) the trial includes. Map which shared scenarios can be run without sales approval. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial Sector pages offer online free tests; Enterprise evaluation scope is not published. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - A permanent business free tier, trial duration/card requirements, Enterprise SSO/admin/connector coverage and fixture access are not established. - free_tier — Is there a free tier for business use? [unknown]: Is there a free tier for business use? Unknown. A permanent business free tier, trial duration/card requirements, Enterprise SSO/admin/connector coverage and fixture access are not established. - self_serve_trial — Is a self-serve business trial available? [partial]: Standard/Premium pages link free testing through Stripe and AGB 3.1 allows self-registration. Trial duration, card requirement and quotas were not verified. Sources: swissgpt:swissgpt:university, swissgpt:swissgpt:terms. - trial_scope — Do trials include SSO, connectors and admin features? [unknown]: Do trials include SSO, connectors and admin features? Unknown. A permanent business free tier, trial duration/card requirements, Enterprise SSO/admin/connector coverage and fixture access are not established. - sales_gate — Does meaningful evaluation require sales contact? [partial]: Individual plans can be purchased by self-registration; the Enterprise integration offer routes to consultation. No account or sales contact was used. Sources: swissgpt:swissgpt:university, swissgpt:swissgpt:terms. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:university, swissgpt:swissgpt:terms #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Teams SaaS has a self-serve trial; self-host evaluation is a separate sales-gated route. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=TypingMind Teams SaaS; Starter, Growth and Professional prerequisites stated per answer; Business quote not obtained; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=multi-tenant-saas; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - free_tier — Is there a free tier for business use? [partial]: The cloud offer advertises a time-limited trial rather than a documented permanent free business tier. Personal free-use rights do not supply Teams governance. Sources: typingmind:typingmind:pricing, typingmind:typingmind:typingmind-vs-typingmind-team. - self_serve_trial — Is a self-serve business trial available? [supported]: The cloud-vs-self-host comparison offers a 14-day self-serve cloud trial with no upfront payment. Sources: typingmind:typingmind:cloud-host-vs-self-host. - trial_scope — Do trials include SSO, connectors and admin features? [partial]: The trial provides cloud evaluation; exact trial entitlements for SSO, SCIM, connectors and API quotas were not independently established. Sources: typingmind:typingmind:cloud-host-vs-self-host. - sales_gate — Does meaningful evaluation require sales contact? [supported]: Cloud starts self-serve; Business and annual self-host require a quote, and self-host has no free trial. Sources: typingmind:typingmind:pricing, typingmind:typingmind:selfhost, typingmind:typingmind:cloud-host-vs-self-host. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:pricing, typingmind:typingmind:typingmind-vs-typingmind-team, typingmind:typingmind:cloud-host-vs-self-host, typingmind:typingmind:selfhost ### VEN-04: Contract terms and pricing transparency Judgment rule: Record minimum seats or commitment, contract length, monthly versus annual billing, price-change rules, cancellation and downgrade terms, and invoicing in CHF/EUR. Classify pricing as transparent, partially specified, quote-only or unknown, naming the missing fields. Quote-only pricing is a disclosed constraint. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial Small plans renew monthly or yearly; Enterprise starts at 10 licences with quoted terms. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - Seat-reduction/downgrade timing, future price-change rules, EUR invoicing and complete Enterprise commitments are not specified. - minimum_commitment — What seat or spend minimum applies? [supported]: Standard/Premium cover up to 9 individual licences; Enterprise starts at 10. Sources: swissgpt:swissgpt:university. - contract_length — What contract lengths are offered? [supported]: AGB 3.1 defaults to one calendar month or twelve months with renewal for the same term; an order may set other terms. Sources: swissgpt:swissgpt:terms. - cancellation — How can the contract be cancelled or reduced? [partial]: Either party may cancel for term end through the platform where available or support; prepaid fees are not refunded. Downgrade rules are not specified. Sources: swissgpt:swissgpt:terms. - local_invoicing — Can it be invoiced in CHF or EUR from a European entity? [partial]: Prices are displayed in CHF and AlpineAI AG is Swiss. Standard offers use credit card; invoices can be offered case by case with 14-day payment. EUR invoicing is unverified. Sources: swissgpt:swissgpt:terms, swissgpt:swissgpt:public-pricing. - transparency_class — Is pricing transparent, partial, quote-only or unknown? [supported]: Partially specified individual-sector prices; quote-only Enterprise from 10 licences. Public prices do not establish tax or all usage/service charges. Sources: swissgpt:swissgpt:university, swissgpt:swissgpt:public-pricing, swissgpt:swissgpt:health-pricing. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:university, swissgpt:swissgpt:terms, swissgpt:swissgpt:public-pricing, swissgpt:swissgpt:health-pricing #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Cloud supports monthly/annual subscriptions; self-host is annual and commercial. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel, Commercial self-hosted TypingMind Teams server; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams Starter/Growth/Professional monthly or annual SaaS; quoted Business and separate annual commercial Teams self-host; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=not specified; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - minimum_commitment — What seat or spend minimum applies? [supported]: Cloud plan cards and self-host licence include five users. Business advertises a USD18,700/year starting quote; no all-scenario minimum was established. Sources: typingmind:typingmind:pricing, typingmind:typingmind:selfhost. - contract_length — What contract lengths are offered? [supported]: Cloud-host vs Self-host explicitly describes monthly SaaS and annual quoted self-host subscriptions; the current SaaS landing also offers annual billing. Teams Terms specify renewal of the agreed term. Sources: typingmind:typingmind:cloud-host-vs-self-host, typingmind:typingmind:pricing, typingmind:typingmind:terms-of-service. - cancellation — How can the contract be cancelled or reduced? [partial]: Cloud billing settings manage subscriptions; Terms make fixed-term Teams/self-host fees generally nonrefundable. Exact mid-term seat-reduction proration and notice deadline were not established. Sources: typingmind:typingmind:manage-your-subscription, typingmind:typingmind:terms-of-service. - local_invoicing — Can it be invoiced in CHF or EUR from a European entity? [unknown]: Can it be invoiced in CHF or EUR from a European entity? CHF/EUR invoicing from a European entity, tax handling and bespoke contract negotiation were not established. - transparency_class — Is pricing transparent, partial, quote-only or unknown? [supported]: Partial: public small-plan SaaS seat pricing with separate API bills, quote-only Business and self-host, and unpriced optional support/knowledge services. Sources: typingmind:typingmind:pricing, typingmind:typingmind:selfhost. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:pricing, typingmind:typingmind:selfhost, typingmind:typingmind:cloud-host-vs-self-host, typingmind:typingmind:terms-of-service, typingmind:typingmind:manage-your-subscription ### VEN-05: Source licence Judgment rule: Record licence per component (client, server, connectors, hosted extensions) and revision. Distinguish OSI open source, source-available with restrictions (branding clauses, user caps), and closed. Repository visibility is not a licence classification. Licences change between releases: record the current terms, and never carry a fork's historical licence over to the upstream product or the reverse. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial Terms grant a contractual right to use a proprietary platform. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - Component licences, branding or user-count restrictions beyond purchased licences, and a separate enterprise source edition are not published. No public implementation repository was identified. - licence_type — Which licence applies to each component? [supported]: AGB 4.1-4.3 grants non-exclusive, non-transferable access during the contract and reserves platform software, algorithms, documentation and branding rights to AlpineAI. Open-model use is separate. Sources: swissgpt:swissgpt:terms. - branding_restrictions — Do licence terms restrict branding or user counts? [unknown]: Do licence terms restrict branding or user counts? Unknown. Component licences, branding or user-count restrictions beyond purchased licences, and a separate enterprise source edition are not published. No public implementation repository was identified. - enterprise_features_closed — Are enterprise features in a separate closed edition? [unknown]: Are enterprise features in a separate closed edition? Unknown. Component licences, branding or user-count restrictions beyond purchased licences, and a separate enterprise source edition are not published. No public implementation repository was identified. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:terms #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial TypingMind is commercial proprietary software; self-host access is not an open-source licence. Availability: generally-available Delivery: native Scope: surfaces=TypingMind personal browser app and PWA, TypingMind Teams SaaS chat and Admin Panel, Commercial self-hosted TypingMind Teams server; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Commercial personal licence, separate bulk pack, Teams SaaS Starter/Growth/Professional and separate commercial Teams self-host licence; personal user/device allowance unverified; add_ons=Customer-paid model APIs; optional knowledge-base capacity and third-party plugin services; deployment_mode=not specified; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - The personal user/device allowance is unknown. No five-user or one-user allowance is published as a reconciled finding; Teams and annual self-host seat rights remain separate. - licence_type — Which licence applies to each component? [supported]: The personal app and Teams operate under commercial Terms/EULA. Self-host provides compiled/minified code in a private repository with no analysis, modification, resale, sharing or redistribution rights; white-box review is by arrangement. Sources: typingmind:typingmind:end-user-license-agreement, typingmind:typingmind:terms-of-service, typingmind:typingmind:selfhost. - branding_restrictions — Do licence terms restrict branding or user counts? [partial]: Teams SaaS provides custom branding; the separate commercial Teams self-host licence includes one instance with five seats. Personal user/device allowance remains unknown because the generic five-device-or-user comparison conflicts with the one-user Terms. The separate bulk pack does not establish the default personal allowance. Sources: typingmind:typingmind:custom-branding, typingmind:typingmind:r2-plan-matrix, typingmind:typingmind:selfhost, typingmind:typingmind:terms-of-service, typingmind:typingmind:r2-personal-bulk-licence. - enterprise_features_closed — Are enterprise features in a separate closed edition? [supported]: Teams administration is part of the commercial service; SSO/SCIM, analytics, roles, logs and API are in Professional rather than an open-source core edition. Sources: typingmind:typingmind:compare-typingmind-team-plans. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:end-user-license-agreement, typingmind:typingmind:terms-of-service, typingmind:typingmind:selfhost, typingmind:typingmind:custom-branding, typingmind:typingmind:r2-plan-matrix, typingmind:typingmind:r2-personal-bulk-licence, typingmind:typingmind:compare-typingmind-team-plans ### VEN-06: Data export and exit Judgment rule: Record admin export of chats, assistants, knowledge bases, prompts and audit logs, formats, export completeness, user self-export, and what survives contract end. Run an admin export on a trial. Exported chats alone do not make assistants and knowledge portable. #### AlpineAI / SwissGPT Verification status: No public information found (not-published) Support: not-published We found no public information on this. A missing finding is not evidence that the product lacks the capability. #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Recorded Teams chats have a JSON API export; generic app exports do not prove full organizational portability. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams REST API, TypingMind Teams SaaS chat and Admin Panel, TypingMind personal browser app and PWA; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams Professional recorded-chat API/export; separate personal app JSON backups (feature licence gate not mapped); add_ons=Professional chat recording enabled before export; authorized API key; deployment_mode=not specified; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Documented findings only; unestablished atomic questions remain explicit gaps. - chat_export — Can admins export all chats? [partial]: Professional includes chat logs/data export and API integration. Authorized admins export recorded conversation JSON via the API, last100 matching conversations/request plus date filters; complete tenant history and pagination were not tested. Sources: typingmind:typingmind:export-chat-logs, typingmind:typingmind:r2-plan-matrix. - assistant_export — Can assistants and prompts be exported in a reusable format? [partial]: The personal-app backup guide exports user agents, prompts/plugins/model settings as JSON. That is not evidence of bulk centrally managed Teams asset export; its personal feature gate is not mapped here. Sources: typingmind:typingmind:export-import-data. - knowledge_export — Can knowledge bases and their sources be exported? [unknown]: Can knowledge bases and their sources be exported? A bulk reusable organization KB/source export and complete centrally managed agent/prompt export were not established. - export_format — Are formats documented and machine-readable? [supported]: Professional recorded-chat API responses are JSON; the separate personal backup/import also uses JSON. Formats do not establish complete organization portability. Sources: typingmind:typingmind:export-chat-logs, typingmind:typingmind:export-import-data, typingmind:typingmind:r2-plan-matrix. - post_termination_window — How long can data be retrieved after termination? [partial]: The DPA allows deletion/return at termination but does not publish a fixed retrieval grace period. Terms allow content deletion after termination. Sources: typingmind:typingmind:data-processing-agreement, typingmind:typingmind:terms-of-service. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:export-chat-logs, typingmind:typingmind:r2-plan-matrix, typingmind:typingmind:export-import-data, typingmind:typingmind:data-processing-agreement, typingmind:typingmind:terms-of-service ### VEN-07: Support and service levels Judgment rule: Record support channels, languages (German, French, Italian), hours and time zone, local or regional presence, documented response times, uptime SLA and credits, public status page and incident history. Popularity is not a reliability score. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial Support uses a service portal and email fallback; terms promise 99% availability. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - DE/FR/IT support guarantees, support hours, response times, SLA credits and an official public status/incident history were not located. - support_channels — Which support channels are available on each plan? [supported]: Service Desk/email is listed in sector plans. AGB 9.2 requires support-portal tickets, with email when the portal is technically unreachable. Sources: swissgpt:swissgpt:university, swissgpt:swissgpt:terms. - support_languages — Is support available in German, French or Italian? [unknown]: Is support available in German, French or Italian? Unknown. DE/FR/IT support guarantees, support hours, response times, SLA credits and an official public status/incident history were not located. - regional_presence — Is there a DACH or Swiss entity, team or partner? [supported]: AlpineAI AG lists a Davos, Switzerland address. Sources: swissgpt:swissgpt:imprint. - uptime_sla — Is there an uptime SLA with credits? [partial]: AGB 9.1 promises 99% availability during the licence term with planned/urgent maintenance exceptions. No credit schedule or response-time commitment was located. Sources: swissgpt:swissgpt:terms. - status_page — Is there a public status page with history? [unknown]: Is there a public status page with history? Unknown. DE/FR/IT support guarantees, support hours, response times, SLA credits and an official public status/incident history were not located. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:university, swissgpt:swissgpt:terms, swissgpt:swissgpt:imprint #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Email support is documented; self-host SLA is optional and public incident data was inaccessible. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel, Commercial self-hosted TypingMind Teams server; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams Starter/Growth/Professional public email support/material; separate quoted annual Teams self-host Enterprise Support add-on; add_ons=Commercial Teams self-host Enterprise Support (SLA/SLO/dedicated queue), separately quoted; deployment_mode=not specified; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Self-host support package is distinct from SaaS support/onboarding; no public numerical uptime/credit or adoption-service tariff established. - support_channels — Which support channels are available on each plan? [supported]: Cloud plan comparison lists email support. Self-host dedicated support/SLA is a separate option; exact Business response targets were not established. Sources: typingmind:typingmind:compare-typingmind-team-plans, typingmind:typingmind:selfhost. - support_languages — Is support available in German, French or Italian? [unknown]: Is support available in German, French or Italian? No German/French/Italian support commitment, DACH/Swiss support entity or public credit-bearing SLA terms were retrieved. - regional_presence — Is there a DACH or Swiss entity, team or partner? [unknown]: Is there a DACH or Swiss entity, team or partner? No German/French/Italian support commitment, DACH/Swiss support entity or public credit-bearing SLA terms were retrieved. - uptime_sla — Is there an uptime SLA with credits? [partial]: Self-host marketing offers a paid SLA/SLO/dedicated-support add-on; a public uptime percentage, credit schedule or standard SaaS SLA was not retrieved. Sources: typingmind:typingmind:selfhost. - status_page — Is there a public status page with history? [partial]: A public status site loads but returned an error fetching status data in this pass, so history and measured availability are unknown. Sources: typingmind:typingmind:status. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:compare-typingmind-team-plans, typingmind:typingmind:selfhost, typingmind:typingmind:status, typingmind:typingmind:pricing, typingmind:typingmind:cloud-host-vs-self-host ### VEN-08: Onboarding, training and adoption Judgment rule: Record vendor-led onboarding, admin training, end-user training (live, e-learning, languages), use-case libraries, champion programmes, change-management materials and their cost. Distinguish included services from paid professional services and partner offerings. #### AlpineAI / SwissGPT Verification status: Agent cross-checked (agent-cross-checked) Support: partial Sector offers describe guided rollout, workshops and a six-lesson e-learning course. Availability: generally-available Delivery: native Scope: surfaces=AlpineAI platform browser client; version=not specified; plan=Standard and Premium, up to 9 licences; Enterprise from 10 licences. Current healthcare, public-administration and university offer.; add_ons=none stated; deployment_mode=not specified; region=Switzerland infrastructure; contract permits CH/EEA and conditional third-country processing; optional external web search; os=not specified. Limitations: - Champion programmes, training languages, mandatory onboarding scope and separate professional-service charges are not published. - onboarding_service — Is guided onboarding included? [partial]: University and healthcare pages offer implementation help and onsite team training; inclusion in licence price is not stated. Sources: swissgpt:swissgpt:university, swissgpt:swissgpt:health-pricing. - user_training — Is end-user training offered, in which languages? [partial]: Six e-learning lessons and prompting/use-case workshops are offered; delivery languages are not specified. Sources: swissgpt:swissgpt:product, swissgpt:swissgpt:university. - use_case_library — Are role-based use cases and templates provided? [supported]: Sector pages describe university, authority and healthcare use cases; Premium adds preconfigured assistants. Sources: swissgpt:swissgpt:university, swissgpt:swissgpt:product. - adoption_programme — Is there a champion or adoption programme? [unknown]: Is there a champion or adoption programme? Unknown. Champion programmes, training languages, mandatory onboarding scope and separate professional-service charges are not published. - service_cost — Which services cost extra? [unknown]: Which services cost extra? Unknown. Champion programmes, training languages, mandatory onboarding scope and separate professional-service charges are not published. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: swissgpt:swissgpt:university, swissgpt:swissgpt:health-pricing, swissgpt:swissgpt:product #### TypingMind / TypingMind Teams Verification status: Agent cross-checked (agent-cross-checked) Support: partial Public setup guides and role examples exist; managed adoption services are not fully specified. Availability: generally-available Delivery: native Scope: surfaces=TypingMind Teams SaaS chat and Admin Panel, Commercial self-hosted TypingMind Teams server; version=Public documentation and changelog retrieved 2026-10-21; no deployed tenant observed; plan=Teams Starter/Growth/Professional public email support/material; separate quoted annual Teams self-host Enterprise Support add-on; add_ons=Commercial Teams self-host Enterprise Support (SLA/SLO/dedicated queue), separately quoted; deployment_mode=not specified; region=US or opt-in EU SaaS storage; inference and tools depend on customer-selected providers; os=not specified. Limitations: - Self-host support package is distinct from SaaS support/onboarding; no public numerical uptime/credit or adoption-service tariff established. - onboarding_service — Is guided onboarding included? [unknown]: Is guided onboarding included? No included guided-onboarding scope, multilingual training commitment or champion/adoption programme was established. - user_training — Is end-user training offered, in which languages? [partial]: Public documentation teaches setup, use, agents and workflows. A contracted live user-training service and languages were not established. Sources: typingmind:typingmind:typingmind-vs-typingmind-team, typingmind:typingmind:overview. - use_case_library — Are role-based use cases and templates provided? [supported]: The documentation provides marketing ad creation, meeting-summary and business automation examples and role/permission examples. Sources: typingmind:typingmind:ad-creative-generator, typingmind:typingmind:meeting-summarizer, typingmind:typingmind:overview, typingmind:typingmind:roles-and-permissions. - adoption_programme — Is there a champion or adoption programme? [unknown]: Is there a champion or adoption programme? No included guided-onboarding scope, multilingual training commitment or champion/adoption programme was established. - service_cost — Which services cost extra? [partial]: Self-host dedicated support/SLA is an optional extra; onboarding/training/adoption-program prices were not established. Sources: typingmind:typingmind:selfhost. Checked 2026-10-21; agent cross-checked 2026-10-21; check again by 2026-12-20. Sources: typingmind:typingmind:typingmind-vs-typingmind-team, typingmind:typingmind:overview, typingmind:typingmind:ad-creative-generator, typingmind:typingmind:meeting-summarizer, typingmind:typingmind:roles-and-permissions, typingmind:typingmind:selfhost, typingmind:typingmind:pricing, typingmind:typingmind:cloud-host-vs-self-host ## Sources ### swissgpt:swissgpt:product AlpineAI product and FAQ. Describes browser and mobile access, Mistral Large 3 on AlpineAI hardware, document assistants, Outlook, Listener and Enterprise integrations. - URL: https://alpineai.swiss/produkt/ - Locator: Chat; Assistants; Listener; FAQs on models, Outlook, licences - Evidence kind: vendor-claimed - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:set-up-api-keys Set up API Keys. Teams admins preset customer provider keys; end users cannot enter their own key in the chat interface. - URL: https://docs.typingmind.com/typingmind-team/getting-started/set-up-api-keys.md - Locator: Set up API Keys - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:manage-ai-models Manage AI Models. The app has a per-chat model selector and configurable model endpoints; provider examples are not a complete current catalogue. - URL: https://docs.typingmind.com/manage-and-connect-ai-models/manage-ai-models.md - Locator: Manage AI Models - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:release TypingMind changelog. Dated releases include Teams memory administration, new models, browser-local code and office outputs, and GPT Image editing. - URL: https://feedback.typingmind.com/changelog/ - Locator: 2026-10-14 Memory for Teams and new models; 2026-10-13 Toolbox; 2026-10-02 image editing - Evidence kind: documented - Source date: 2026-10-14 - Retrieved: 2026-10-21 ### typingmind:typingmind:restrict-model-access Restrict Model access. Teams admins restrict available models by user group. - URL: https://docs.typingmind.com/typingmind-team/branding-and-customizations/restrict-model-access.md - Locator: Restrict Model access - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:chat-with-documents Chat with documents. Uploads accept PDF, Markdown, CSV, XLSX, TXT, DOCX, JSON and HTML, with document text added as chat context. - URL: https://docs.typingmind.com/upload-and-chat-with-files/chat-with-documents.md - Locator: Chat with documents - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:voice Voice input. Browser speech recognition or customer OpenAI, Groq and Azure transcription keys enable dictation; uploaded Whisper recordings have a 25MB limit. - URL: https://docs.typingmind.com/voice-input - Locator: Transcription models and Teams configuration - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:azure-openai Azure OpenAI. A customer Azure OpenAI endpoint and deployment can be configured. - URL: https://docs.typingmind.com/manage-and-connect-ai-models/azure-openai.md - Locator: Azure OpenAI - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:pricing TypingMind Teams pricing. Annual selected prices display USD83/166/249 per month for Starter/Growth/Professional, five users, USD8 extra seats; Business starts USD18,700/year. - URL: https://custom.typingmind.com/pricing - Locator: Annual selector, plan cards, Add-ons and Business card - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:get-started-with-self-host-installation-instruct Self-host installation instructions. A NodeJS service needs MySQL 8, optional PostgreSQL 16 pgvector KB, OpenAI embeddings and S3-compatible blobs; private endpoints are blocked by default. - URL: https://docs.typingmind.com/typingmind-team/self-host-guides/get-started-with-self-host-installation-instruct.md - Locator: Self-host installation instructions - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:toolbox Toolbox. A free built-in plugin executes Python and JavaScript locally in the browser with network disabled by default and creates Office/PDF files and charts. - URL: https://docs.typingmind.com/plugins/toolbox - Locator: Local execution, network control and file capabilities - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:r2-teams-custom-models Teams API Keys Setup. Teams-specific instructions name Azure OpenAI, Hugging Face and Ollama custom-model setup in the Admin Panel; this is distinct from personal local-server guides. - URL: https://docs.typingmind.com/typingmind-team/getting-started/set-up-api-keys.md - Locator: Use with other models; Add Custom Model in Admin Panel - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:r2-plan-matrix Compare Teams Plans. All SaaS tiers list custom models and shared resources; Growth adds RAG/projects; Professional adds roles, restrictions, logs and external API integration. - URL: https://docs.typingmind.com/typingmind-team/getting-started/compare-typingmind-team-plans.md - Locator: Full Feature Comparison - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:r3-custom-admin-bedrock AWS Bedrock Anthropic. The Custom Admin route configures Claude via Bedrock with customer AWS access/secret keys, region and approved model access; current model and commercial tier coverage are not enumerated. - URL: https://docs.typingmind.com/manage-and-connect-ai-models/aws-bedrock-anthropic - Locator: Custom Admin restriction and steps 2-4, including custom.typingmind.com/manage - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:ollama Ollama. A customer Ollama model endpoint can be added; network accessibility must be configured. - URL: https://docs.typingmind.com/manage-and-connect-ai-models/ollama.md - Locator: Ollama - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:lmstudio LM Studio. Customer local LM Studio can supply an OpenAI-compatible model endpoint. - URL: https://docs.typingmind.com/manage-and-connect-ai-models/lmstudio.md - Locator: LM Studio - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:cloud-host-vs-self-host Cloud host vs self host. SaaS has a self-serve 14-day trial; self-host has no free trial and carries a quoted annual licence and update/support conditions. - URL: https://docs.typingmind.com/typingmind-team/getting-started/cloud-host-vs-self-host.md - Locator: Cloud host vs self host - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:usage-limits Usage limits. Reusable limit groups cover messages, characters and tokens by period, applicable globally or to users, groups, models, agents and plugins. - URL: https://docs.typingmind.com/typingmind-team/user-management/usage-limits.md - Locator: Usage limits - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:connect-knowledge-base-to-your-ai-agents Connect Knowledge Base to your AI Agents. Tagged knowledge is assigned to agents through Knowledge Base Access. - URL: https://docs.typingmind.com/typingmind-team/branding-and-customizations/connect-knowledge-base-to-your-ai-agents.md - Locator: Connect Knowledge Base to your AI Agents - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:roles-and-permissions Roles and Permissions. Owners delegate granular Admin Panel permissions through custom roles; group-role assignment is described as coming soon. - URL: https://docs.typingmind.com/typingmind-team/user-management/roles-and-permissions.md - Locator: Roles and Permissions - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### swissgpt:swissgpt:university AlpineAI universities. Lists licence limits and Enterprise threshold, university integration process and onboarding. - URL: https://alpineai.swiss/en/universities/ - Locator: Prices; Integration & Licenses; Features & Languages - Evidence kind: vendor-claimed - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:how-knowledge-base-works-in-typingmind-team How Knowledge Base works in TypingMind Team. The early-access KB uses RAG, about 3,000-word chunks and up to five retrieved chunks, 50MB uploads and daily source refresh. - URL: https://docs.typingmind.com/typingmind-team/branding-and-customizations/how-knowledge-base-works-in-typingmind-team.md - Locator: Early Access; indexing; limits and refresh - Evidence kind: documented - Source date: 2025-11-10 - Retrieved: 2026-10-21 ### typingmind:typingmind:training-files-vs-knowledge-base Training Files vs Knowledge Base. Agent training-file text is injected into system context; the separate KB retrieves chunks and advertises multiple source integrations. - URL: https://docs.typingmind.com/typingmind-team/branding-and-customizations/training-files-vs-knowledge-base.md - Locator: Training Files vs Knowledge Base - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:get-started-with-knowledge-base Get started with Knowledge Base. Admins add documents or connected sources, tag data, bind tags to agents and optionally show clickable references. - URL: https://docs.typingmind.com/typingmind-team/branding-and-customizations/get-started-with-knowledge-base.md - Locator: Get started with Knowledge Base - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:restrict-ai-agents-access Restrict AI Agents access. Agent access and usage can be limited for groups. - URL: https://docs.typingmind.com/typingmind-team/branding-and-customizations/restrict-ai-agents-access.md - Locator: Restrict AI Agents access - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### swissgpt:swissgpt:product-20261026 AlpineAI product and FAQ. Describes browser and mobile access, Mistral Large 3 on AlpineAI hardware, document assistants, Outlook, Listener and Enterprise integrations. - URL: https://alpineai.swiss/produkt/ - Locator: Chat; Assistants; Listener; FAQs on models, Outlook, licences - Evidence kind: vendor-claimed - Source date: Not stated - Retrieved: 2026-10-26 ### swissgpt:swissgpt:security-20261026 AlpineAI security and compliance. Separates RIZ application hosting from Green and RZO inference and describes encryption, access restrictions and external-search opt-in. - URL: https://alpineai.swiss/sicherheit/ - Locator: Infrastructure; isolation; encryption; security FAQ - Evidence kind: vendor-claimed - Source date: Not stated - Retrieved: 2026-10-26 ### swissgpt:swissgpt:public-pricing-20261026 AlpineAI authorities and institutions. Monthly billing is CHF 79 Standard or CHF 89 Premium per user; annual billing displays CHF 74 or CHF 80 per month; Enterprise is quoted from 10 licences. - URL: https://alpineai.swiss/behoerden/ - Locator: Prices; pricing-toggle applyMode(monthly/yearly) and billing labels in delivered HTML - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-26 ### swissgpt:swissgpt:health-pricing-20261026 AlpineAI healthcare. Healthcare monthly billing is CHF 89 Standard or CHF 99 Premium per user; annual billing displays CHF 80 or CHF 90 per month. - URL: https://alpineai.swiss/gesundheitswesen/ - Locator: Prices; pricing-toggle applyMode(monthly/yearly) and billing labels in delivered HTML - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-26 ### typingmind:typingmind:connect-with-your-notion-pages Connect with Notion pages. Admins select Notion pages/databases for an instance, with initial sync taking a few minutes. - URL: https://docs.typingmind.com/typingmind-team/branding-and-customizations/connect-with-your-notion-pages.md - Locator: Connect with Notion pages - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:setup-external-connectors-for-knowledge-base Self-host external knowledge connectors. Optional self-host KB external connectors need separate OAuth configuration and Firecrawl for web scraping. - URL: https://docs.typingmind.com/typingmind-team/self-host-guides/setup-external-connectors-for-knowledge-base.md - Locator: Self-host external knowledge connectors - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:overview LLM Automated Workflows overview. JavaScript, HTTP and MCP plugins can perform multi-step business actions; example flows are not evidence of a scheduler. - URL: https://docs.typingmind.com/llm-automated-workflows/overview.md - Locator: LLM Automated Workflows overview - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:mcp MCP in TypingMind. The client connects remote HTTP or local-command MCP servers, supports OAuth and individual tool controls, and separates Teams sessions per user/chat. - URL: https://docs.typingmind.com/model-context-protocol-(mcp)-in-typingmind - Locator: Remote and local connectors; Teams sessions; tool control - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:eu-data-center EU data center. EU instances use Vercel and AWS EU storage but model API calls may leave the EU. - URL: https://docs.typingmind.com/typingmind-team/account-management/eu-data-center.md - Locator: EU data center - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:system-prompts-on-admin-panel System Prompts on Admin Panel. Global system instructions, few-shot examples and post-message reminders apply across chats and agents. - URL: https://docs.typingmind.com/typingmind-team/branding-and-customizations/system-prompts-on-admin-panel.md - Locator: System Prompts on Admin Panel - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:memory Memory. Users can enable memory globally or for an agent/project and inspect saved memories. - URL: https://docs.typingmind.com/memory.md - Locator: Memory - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:build-a-shared-ai-agent-collection Build a Shared AI Agent collection. Admins create an agent collection exposed in the team chat interface and control group access. - URL: https://docs.typingmind.com/typingmind-team/branding-and-customizations/build-a-shared-ai-agent-collection.md - Locator: Build a Shared AI Agent collection - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:build-a-shared-prompt-library Build a Shared Prompt Library. Admins can publish reusable prompts to the team interface. - URL: https://docs.typingmind.com/typingmind-team/branding-and-customizations/build-a-shared-prompt-library.md - Locator: Build a Shared Prompt Library - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:restrict-plugin-access Restrict Plugin access. Plugins can be made available or unavailable to specified user groups. - URL: https://docs.typingmind.com/typingmind-team/branding-and-customizations/restrict-plugin-access.md - Locator: Restrict Plugin access - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:deep-research Deep Research plugin. A multi-step search/read/report plugin requires customer SerpAPI and Firecrawl keys and has a two-minute execution limit. - URL: https://docs.typingmind.com/plugins/deep-research - Locator: Requirements and operation - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:image GPT Image 2.5 plugin. The plugin generates and edits images using the customer OpenAI API route. - URL: https://docs.typingmind.com/plugins/gpt-image-2-5 - Locator: Image generation, editing and API key - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### swissgpt:swissgpt:terms AlpineAI general terms. Defines cloud licensing, no-training promise, web-search consent and quotas, subscription cancellation, 99 percent availability and Swiss law. - URL: https://alpineai.swiss/agb/ - Locator: Sections 2.1-2.5, 3, 4, 7.5, 9 and 11 - Evidence kind: documented - Source date: 2026-05-25 - Retrieved: 2026-10-21 ### swissgpt:swissgpt:dpa AlpineAI data processing agreement. Sets CH/EEA processing with a third-country exception, anonymised-data use, deletion and backup exception, safeguards and a dated subprocessor list. - URL: https://alpineai.swiss/wp-content/uploads/2026/07/AlpineAI-AVV_zu_GTC_2026-07-10.pdf - Locator: Pages 1-4 sections 1.5, 2.4, 7, 9 and 11; pages 6-7 Annexes 2-3 - Evidence kind: documented - Source date: 2026-07-10 - Retrieved: 2026-10-21 ### typingmind:typingmind:api-integration-overview API integration overview. REST chat/agent API uses API-key-associated api-users subject to user restrictions and usage limits. - URL: https://docs.typingmind.com/typingmind-team/integrations/api-integration-overview.md - Locator: API integration overview - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### swissgpt:swissgpt:review2-product AlpineAI product and FAQ. The product page assigns Standard/Premium to assistants, not autonomous agents; it explicitly assigns custom integrations to Enterprise. - URL: https://alpineai.swiss/produkt/ - Locator: Integrations; Assistants & Agents; Operations & Licences - Evidence kind: vendor-claimed - Source date: Not stated - Retrieved: 2026-10-21 ### swissgpt:swissgpt:review2-api-access AlpineAI API platform. Public API portal redirects to authentication and states that access is limited with gradual account activation; no authenticated endpoint/entitlement documentation was accessed. - URL: https://platform.prod.alpineai.ch/ - Locator: Authentication landing page; limited access notice - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:typingmind-vs-typingmind-team TypingMind vs TypingMind Team. The personal app and Teams with an Admin Panel, managed users and centrally preset keys are separate offerings. - URL: https://docs.typingmind.com/quickstart/typingmind-vs-typingmind-team.md - Locator: TypingMind vs TypingMind Team - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:install Install TypingMind app. Installation uses browser PWA technology; browser storage remains separate and this does not establish native app-store clients. - URL: https://docs.typingmind.com/install-typingmind-app - Locator: PWA installation on Windows, Mac, iOS and Android - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:r2-personal-bulk-licence TypingMind personal licence checkout. Personal licences normally cover five devices for one user; a separate shareable Bulk License is offered, without establishing a Teams Admin Panel entitlement. - URL: https://checkout.typingmind.com/buy - Locator: How many devices/users? and Team/Bulk License FAQ; self-host FAQ - Evidence kind: vendor-claimed - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:ad-creative-generator Ad creative generator workflow. Public examples include marketing automation built with AI and plugins. - URL: https://docs.typingmind.com/llm-automated-workflows/ad-creative-generator.md - Locator: Ad creative generator workflow - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:terms-of-service Terms of Service. Typing Technology Co Ltd contracts under Singapore law and business SIAC arbitration; Teams renews and self-host code has restricted licence rights. - URL: https://docs.typingmind.com/security-and-compliance/terms-of-service.md - Locator: Sections 1, 4-6, 13 and 16 - Evidence kind: documented - Source date: 2026-09-17 - Retrieved: 2026-10-21 ### typingmind:typingmind:tts Text to speech. Answers can be read aloud using browser speech, ElevenLabs or OpenAI TTS. - URL: https://docs.typingmind.com/use-text-to-speech - Locator: Browser, ElevenLabs and OpenAI options - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:meeting-summarizer Meeting summarizer workflow. The guide offers an audio-transcription and summarisation workflow, not a demonstrated meeting bot. - URL: https://docs.typingmind.com/llm-automated-workflows/meeting-summarizer.md - Locator: Meeting summarizer workflow - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:compare-typingmind-team-plans Compare TypingMind Team plans. The comparison lists monthly USD99/199/299, five users plus USD8 additional users, Growth knowledge and Professional identity/governance. - URL: https://docs.typingmind.com/typingmind-team/getting-started/compare-typingmind-team-plans.md - Locator: Compare TypingMind Team plans - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:translations Translation collaboration. A translation project exists, but the retrieved page does not establish completeness of German, French, Italian or Admin Panel translations. - URL: https://docs.typingmind.com/app-translation/translation-collaboration - Locator: Multilingual app translation project - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:sso Single Sign-On (SSO). SAML identity-provider metadata, entity ID and reply URL are configured in Teams. - URL: https://docs.typingmind.com/typingmind-team/user-management/single-sign-on-(sso) - Locator: SAML configuration and multiple connections - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:o-auth-2-0 OAuth 2.0. OAuth/OIDC supports built-in Microsoft and Google logins and JIT seats; revoking IdP authorization alone does not remove the user. - URL: https://docs.typingmind.com/typingmind-team/user-management/o-auth-2-0.md - Locator: OAuth 2.0 - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:data-processing-agreement Data Processing Agreement. The public GDPR processor form includes confidentiality, subprocessor flow-down, SCC transfers and customer-choice deletion or return. - URL: https://docs.typingmind.com/security-and-compliance/data-processing-agreement.md - Locator: Clauses 7.4, 7.8-7.10 and Appendices I-IV - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:scim Directory Sync (SCIM v2). SCIM sync creates and suspends users, updates profiles and assigns corresponding groups from the IdP. - URL: https://docs.typingmind.com/typingmind-team/user-management/directory-sync-(scim-v2) - Locator: Create, update, suspend users and groups - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:invite-team-members Invite team members. Teams admins invite members to the instance and manage seats. - URL: https://docs.typingmind.com/typingmind-team/user-management/invite-team-members.md - Locator: Invite team members - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:create-and-manage-user-groups Create and manage user groups. Groups control access and are usable for limits and analytics. - URL: https://docs.typingmind.com/typingmind-team/user-management/create-and-manage-user-groups.md - Locator: Create and manage user groups - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:chatbot-analytics Chatbot analytics. Reporting covers active users, messages, conversations, estimated tokens and characters, with model/agent/group/user breakdowns and chart downloads. - URL: https://docs.typingmind.com/typingmind-team/analytics/chatbot-analytics.md - Locator: Chatbot analytics - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:view-and-export-user-chat-logs View and export user chat logs. Opt-in Record all chats exposes user prompts and responses to authorized admins and configurable retention permanently deletes old records. - URL: https://docs.typingmind.com/typingmind-team/analytics/view-and-export-user-chat-logs.md - Locator: View and export user chat logs - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:export-chat-logs Export chat logs. Chat logging is off by default; the API exports JSON conversation content and dates, limited to the last 100 matching conversations per request. - URL: https://docs.typingmind.com/typingmind-team/integrations/export-chat-logs.md - Locator: Export chat logs - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### swissgpt:swissgpt:secrecy AlpineAI supplementary agreement for secret data. When signed for legally secret data, requires Swiss plaintext services, restricted access and zero prompt/response retention for AI API services. - URL: https://alpineai.swiss/wp-content/uploads/2026/03/AlpineAI_Zusatzvereinbarung_Geheimnisdaten.2025-03-30-1.pdf - Locator: Printed version 30 March 2026; section 3 on both pages - Evidence kind: documented - Source date: 2026-03-30 - Retrieved: 2026-10-21 ### swissgpt:swissgpt:review2-security AlpineAI security and compliance. Security page describes RIZ application hosting and Green/RZO inference as colocation on AlpineAI-owned/operated hardware, without logical access by site operators. This is a vendor infrastructure disclosure, not an updated DPA schedule. - URL: https://alpineai.swiss/sicherheit/ - Locator: Certified infrastructure; FAQ on storage and processing - Evidence kind: vendor-claimed - Source date: Not stated - Retrieved: 2026-10-21 ### swissgpt:swissgpt:review2-dpa AlpineAI data processing agreement. Re-retrieved July 10 Annex 3 lists RIZ hosting/support and Green pure colocation with no logical data access, plus CH/EU/US operational suppliers; RZO is not listed. No newer public schedule or RZO approval was retrieved. - URL: https://alpineai.swiss/wp-content/uploads/2026/07/AlpineAI-AVV_zu_GTC_2026-07-10.pdf - Locator: Annex 3 page 7, dated 10 July 2026; section 1.5 and 7.3 - Evidence kind: documented - Source date: 2026-07-10 - Retrieved: 2026-10-21 ### typingmind:typingmind:selfhost TypingMind Teams Self Host. Commercial self-host requires an annual quote, includes five users, uses compiled/minified code, checks licence/updates externally and can use customer model endpoints. - URL: https://custom.typingmind.com/self-host - Locator: Technical requirements, licensing, network, updates and pricing FAQ - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:privacy-policy Privacy policy. The policy distinguishes personal local storage from Teams multitenant server storage, and lists inactive unpaid account deletion. - URL: https://docs.typingmind.com/security-and-compliance/privacy-policy.md - Locator: Storage scope, API keys and section on account deletion - Evidence kind: documented - Source date: 2026-09-19 - Retrieved: 2026-10-21 ### typingmind:typingmind:data-usage-and-model-training-policy Data Usage and Model Training Policy. The vendor says it does not train on conversations; customers contract directly with model providers and their retention/training terms vary. - URL: https://docs.typingmind.com/security-and-compliance/data-usage-and-model-training-policy.md - Locator: TypingMind role and provider API comparison - Evidence kind: vendor-claimed - Source date: Not stated - Retrieved: 2026-10-21 ### swissgpt:swissgpt:imprint AlpineAI imprint. Identifies AlpineAI AG at Obere Strasse 22b in Davos, Switzerland. - URL: https://alpineai.swiss/impressum/ - Locator: Address - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### swissgpt:swissgpt:security AlpineAI security and compliance. Separates RIZ application hosting from Green and RZO inference and describes encryption, access restrictions and external-search opt-in. - URL: https://alpineai.swiss/sicherheit/ - Locator: Infrastructure; isolation; encryption; security FAQ - Evidence kind: vendor-claimed - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:r4-soc2-selfhost-claim TypingMind Teams self-host. The self-host marketing page claims SOC 2 Type II; no auditor, reporting period, covered legal entity or report was retrieved. This is a vendor claim rather than verified attestation. - URL: https://custom.typingmind.com/self-host - Locator: Maximum security, SOC 2 Type II badge; updates and licence checks - Evidence kind: vendor-claimed - Source date: Not stated - Retrieved: 2026-10-26 ### typingmind:typingmind:r2-teams-server-route TypingMind Teams pricing FAQ. The Teams-specific FAQ says users do not see the key and OpenAI requests are sent from the Teams server. - URL: https://custom.typingmind.com/pricing - Locator: Do my users see my OpenAI API key? - Evidence kind: vendor-claimed - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:dpa-confidentiality TypingMind DPA confidentiality. Staff have need-to-know confidentiality duties and subprocessor contracts carry substantially the same data-protection obligations; a specific professional-secrecy agreement is not established. - URL: https://docs.typingmind.com/security-and-compliance/data-processing-agreement.md - Locator: 7.4(c),7.8(d)-(e); AppendixII andIII - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### swissgpt:swissgpt:public-pricing AlpineAI authorities and institutions. Monthly billing is CHF 79 Standard or CHF 89 Premium per user; annual billing displays CHF 74 or CHF 80 per month; Enterprise is quoted from 10 licences. - URL: https://alpineai.swiss/behoerden/ - Locator: Prices; pricing-toggle applyMode(monthly/yearly) and billing labels in delivered HTML - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### swissgpt:swissgpt:health-pricing AlpineAI healthcare. Healthcare monthly billing is CHF 89 Standard or CHF 99 Premium per user; annual billing displays CHF 80 or CHF 90 per month. - URL: https://alpineai.swiss/gesundheitswesen/ - Locator: Prices; pricing-toggle applyMode(monthly/yearly) and billing labels in delivered HTML - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### swissgpt:swissgpt:ecb-fx ECB euro foreign exchange reference rates, 20 October 2026. ECB records EUR 1 = USD 1.1215 and CHF 0.9237 on 20 October 2026; derived USD to EUR is 1/1.1215 and CHF to EUR is 1/0.9237. - URL: https://www.ecb.europa.eu/stats/eurofxref/eurofxref-hist.xml - Locator: Cube time="2026-10-20": USD and CHF - Evidence kind: documented - Source date: 2026-10-20 - Retrieved: 2026-10-21 ### typingmind:typingmind:openai-price OpenAI API pricing. gpt-6-luna Standard input is USD0.10 and output USD0.50 per million tokens, excluding cache and optional regional uplifts. - URL: https://developers.openai.com/api/docs/pricing - Locator: Standard short-context gpt-6-luna row - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:openai-model GPT-6 Luna model documentation. Reasoning effort none is the lowest option; the shared cost proxy does not use reasoning tokens. - URL: https://developers.openai.com/api/docs/models/gpt-6-luna - Locator: Reasoning effort and API support - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:fx ECB reference exchange rates. EUR1 equals USD1.1215 and CHF0.9237. - URL: https://www.ecb.europa.eu/stats/eurofxref/eurofxref-hist.xml - Locator: Cube time=2026-10-20 - Evidence kind: documented - Source date: 2026-10-20 - Retrieved: 2026-10-21 ### typingmind:typingmind:manage-your-subscription Manage your subscription. Cloud customers manage the recurring subscription through billing settings. - URL: https://docs.typingmind.com/typingmind-team/account-management/manage-your-subscription.md - Locator: Manage your subscription - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:end-user-license-agreement End User License Agreement. The software is commercially licensed with restrictions on redistribution and modification. - URL: https://docs.typingmind.com/security-and-compliance/end-user-license-agreement.md - Locator: End User License Agreement - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:custom-branding Custom branding. Teams supports instance logo, theme and branded presentation within its commercial licence. - URL: https://docs.typingmind.com/typingmind-team/branding-and-customizations/custom-branding.md - Locator: Custom branding - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:export-import-data Export and import data. User app settings export/import chats, prompts, plugins, folders, agents and model settings as JSON. - URL: https://docs.typingmind.com/cloud-sync-and-backup/export-import-data.md - Locator: Export and import data - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21 ### typingmind:typingmind:status TypingMind status. The status site loaded a status shell but reported an error fetching data; incident history and measured uptime were not retrieved. - URL: https://status.typingmind.ai - Locator: Public status shell - Evidence kind: documented - Source date: Not stated - Retrieved: 2026-10-21