I use more than one coding agent, and couldn't answer simple questions about my own work. What did a session cost? Which model did I use last Wednesday? Where did that fix come from? The answers were scattered across tools with their own dashboards, or none at all.
So I built the one I wanted.
Firekeeper watches the coding agents on your machine and puts it all in one place: sessions, model selection, token usage, and the full conversation history.
It reads what your agents already write to disk.
Use it however you'd like:
- In the terminal. See actively running agents and tend to your own bonfire.
- In the browser. A dashboard with a session list, live updates, and a transcript viewer.
- In the background. A small daemon keeps the dashboard current while you work.
- Just to look around. Don't want a daemon? Run one import script, see your history in the dashboard, and stop there.