Faisty Launches Public Beta: Turn Your Fastmail Inbox Into a Queryable SQL Database
Key Takeaways
- ▸Email becomes a structured, queryable database with full-text and semantic search via embeddings
- ▸AI agents can read inbox data through SQL or MCP without ever receiving Fastmail credentials—a significant privacy and security improvement
- ▸Self-hosted, Supabase-based architecture ensures data ownership; users can export with pg_dump and cancel anytime
Summary
Faisty, a new startup, has launched its public beta service that transforms Fastmail email accounts into queryable SQL databases with AI-powered semantic search. The platform mirrors every message into a PostgreSQL database (users supply their own Supabase project) and adds embeddings for natural language search capabilities. Faisty offers two interface layers—natural language for humans and SQL/MCP for AI agents—allowing tools like Claude, Cursor, and OpenWebUI to read and query email without accessing credentials. The service operates on a live JMAP stream, syncing new messages to the database within a second of arrival. Users can start free with limited features (5,000 email backfill, limited queries) and upgrade to an unlimited tier for full archive access and sub-second live updates. The platform emphasizes data ownership, as all data lives on users' own Supabase instances—Faisty provides only the pipeline.
- Live JMAP stream keeps the database in sync within a second of new mail arrival, eliminating scheduling or refresh overhead
- Designed for the emerging pattern of AI agents reading and acting on email data
Editorial Opinion
Faisty represents a practical approach to the emerging need for AI agents to access personal data like email. By using embeddings and SQL as the common interface, it elegantly bridges the gap between human-readable queries and agent-executable SQL. The emphasis on data ownership through self-hosted Supabase is refreshing—especially as more tools ask for email access. However, the real test will be adoption among both individual users and AI app developers; the product sits at an interesting intersection of productivity tooling and AI infrastructure.
