See exactly what your AI agents did, what they cost, and where they got stuck — without changing a single line of code.
No spam. No credit card. Just early access.
Built for people who run AI agents every day — not teams managing production infrastructure.
See exactly what every session costs. Set daily and weekly budgets. Get alerted before a runaway agent drains your wallet.
Replay any session step by step — what tools were called, what files were touched, and exactly where things went sideways.
Automatically spots when an agent is stuck repeating the same tool calls. Know in seconds, not after $10 of wasted tokens.
Weekly digests of which agents you're using most, your average session cost, and where your token budget is actually going.
Reads your Claude Code session logs directly — no instrumentation, no data leaving your machine, no code changes required.
Works out of the box with Claude Code. Point it at your sessions folder and you're done. No SDK, no API keys, no config.
Claude Code already writes detailed session logs locally. Debrief reads them. That's it.
Download the Mac app. No account required to get started. Your data stays on your machine.
Debrief automatically finds your Claude Code session logs at ~/.claude/projects/. Nothing to configure.
Every session appears instantly — costs, timelines, tool calls, errors. Past sessions included, not just future ones.
Add budget alerts so you get notified before a session goes off the rails. Take back control of your API spend.
Join the waitlist and be first to try Debrief when it launches.
Free to join. No credit card required.