Security & data access
Last updated: October 4, 2026
What Klynt reads
- Slack — new text messages in channels you explicitly pick. No history is imported, only messages sent after you connect.
- Telegram — new text messages in the single project chat the bot was added to.
- Figma — comments and version history of the file you link.
- Linear — issue status changes and comments in the teams you pick.
What Klynt never does
- No DMs or private chats — the Slack app never requests those permissions.
- No messages outside the channels you picked — filtered in code, not just by policy.
- No posting or reacting in Slack on your behalf. (The Telegram bot sends one confirmation message when a chat connects.)
- We never sell your data or share it with third parties.
How your data is stored
Project data lives in a Postgres database (Supabase). Access tokens are encrypted at rest with AES-256-GCM and all traffic is served over TLS. Raw messages are kept so every fact has verifiable evidence behind it — and so you can inspect or delete them.
How AI processes your data
Incoming events are sent to a language model provider (Anthropic by default) which extracts decisions and state changes into structured facts. Your data is not used to train models. Only the message text needed for interpretation leaves the database; what Klynt stores are structured facts linked back to the original events.
Who can see imported data
Only members of the project see its events and facts. People appearing in imported messages are shown as attribution names — they get no access to the project. Nothing imported is visible to other projects or other Klynt customers.
Disconnect and deletion
Disconnect a source in one click from Integrations: the connection is revoked and messages imported from it are deleted. Deleting a project removes everything — events, facts, members and integrations.
Questions?
For security questions or a review, write to hello@klynt.one — we are happy to walk your security team through this document.