Guide
Privacy (plain language)
What stays on your machine — and the few cases that need the network.
- Chat & page analysis run on-device after the model is downloaded.
- Memories, projects, journal, embeddings live in local browser storage (IndexedDB).
- Privane does not send your normal chat transcripts to a hosted LLM API for inference.
When the network is used
- First model download (Hugging Face)
- Research / Ask Agent web search fetches
- License validation
- Uninstall feedback form (only if you submit it)
Full legal details: Privacy Policy.