Where AI appears
Anasa uses generative or embedding models only in features explicitly presented as AI: chat, writing assistance, summaries, tag and task suggestions, related-note suggestions, semantic search, and similar commands.
Processing choices
- Local: Apple Intelligence or Ollama on the user's device.
- Bring your own key: a supported provider selected and paid by the user.
- Managed: requests routed through Anasa and the Vercel AI Gateway to the model configured for that feature.
Model names and availability can change because provider catalogs change. The app should show the active provider and model where that information affects the user's choice.
Human review
AI output is probabilistic. It may invent facts, miss context, reproduce bias, suggest unsafe actions, or format a file incorrectly. Users should review proposed content and actions before relying on them. High-impact decisions should not be delegated to Anasa AI without qualified human review.
Data and training
Anasa does not use note content to train its own models. Managed requests transmit only the context required for the invoked feature and are governed by the AI Privacy disclosure and the underlying providers' terms. Bring-your-own-key requests are governed by the user's provider agreement.
Automated decisions
Anasa does not use AI to make employment, credit, housing, education, insurance, legal, or similarly significant decisions about users. Suggestions inside the app do not have legal or similarly significant effect by themselves.