Features

AI Notes

Notes that AI leaves for itself — so the next session picks up right where the last one left off.

What are AI notes?

When AI finishes a work session, it can leave notes summarizing what it did, what it found, and what needs attention next. These notes are stored in Matchanu and read back automatically in the next session.

Think of it like a logbook — AI hands off to its future self, so nothing important gets lost between sessions. You can also write notes yourself if you want to leave instructions for AI.

Note types

🤝 Handoff

A summary of what was done in this session and what's left to do. Useful at the end of a coding session.

"Implemented user login. Next: add email verification. The users table already has a verified_at column."

⚙️ Implementation

Technical details that aren't obvious from reading the code — workarounds, edge cases, architectural choices made during implementation.

"Using optimistic updates on the kanban board because the API response is slow. Rollback logic is in useKanban.ts line 84."

⚠️ Risk

Something that could go wrong or needs attention before shipping.

"The password reset flow doesn't rate limit — someone could spam reset emails. Needs to be fixed before launch."

🔍 Discovery

Something interesting or unexpected found during the work — useful context for future decisions.

"Found that Supabase auth tokens expire after 1 hour. The refresh logic needs to be tested under load."

Reading AI notes

1Go to your project and open the AI Notes tab
2Notes are sorted by most recent — scroll to see older ones
3Click any note to expand it
4Notes marked risk are highlighted so you don't miss them

💡AI notes are also included in prepare_agent_brief — so when AI loads your project, it reads all notes automatically.