The assistant that answers with its sources.
Kestrel reads your documents, drafts, and threads — then gives you answers you can verify, citing exactly where each one came from. Private by default, fast enough to think alongside.
- Revenue up 22% QoQ, driven by enterprise seats.
- Net burn held flat at $410k/mo.
- Runway: ~14 months at current burn — down from 17 last quarter.
An assistant you can actually trust with the real work.
Most AI tools sound confident and leave you guessing. Kestrel is built the other way around — around evidence, privacy, and speed.
Private by design
Your workspace is encrypted end to end and never used to train models. What you put in stays yours — full stop.
Answers with sources
Every response links back to the exact file, page, or message it came from — so you verify in one click instead of trusting blind.
Lives in your stack
Drive, Slack, Notion, GitHub and more, connected in a click. Kestrel works where your team already works.
Bring Kestrel to your team.
Start free for up to five seats. No card required, connect a source and ask your first question in under a minute.