If you’ve been using Claude Code, you’ve probably noticed that keeping track of what you actually did is… hard.

Everything gets stored in raw JSONL files on your machine. Great for privacy, terrible for readability.

I got tired of grepping through logs to find “that one prompt that worked perfectly,” so I updated Poirot to handle it for me.

Here’s what’s new:

1. Prompt History: Browse every command you’ve ever sent. Grouped by date, fully searchable.
2. AI Session Summaries: See exactly what happened in a session without reading the whole transcript. It pulls the goal, outcome, and friction points directly from Claude’s usage data.
3. Memory Browser: Peek into what Claude “knows” about your projects via the auto-generated memory files.

It’s all local, reading from your existing ~/.claude config.

It’s open source and free.