If you use Claude Code, Codex, Gemini CLI, or any AI coding agent daily, you’ve accumulated hundreds of sessions. Each one contains decisions, debugging breakthroughs, architectural discussions — your best thinking, scattered across session files on your machine.
Spool makes all of that searchable.
The problem
Your past agent sessions are gold. You’ve solved hard problems, explored trade-offs, and built up context that’s invaluable for future work. But there’s no good way to find any of it.
You can’t grep through JSONL files and get useful results. You can’t ask your agent “what did we discuss about caching last month?” because it has no memory across sessions.
How Spool works
Spool watches your session directories in real time. Every conversation becomes searchable the moment it’s written — no manual export, no copy-paste.
It also indexes data from installable connector plugins: your GitHub stars, Twitter bookmarks, Reddit saves, and more. All local, all on your machine.
What’s next: agent-native search
The key insight: your coding agent is already the best search engine you have. It just needs access to your personal data.
We’re building a /spool skill for Claude Code and a standalone CLI so your agent can search your past sessions and pull matching context directly into the current conversation. Ask it to “build on last month’s auth discussion” and it will actually be able to.
Try it
curl -fsSL https://spool.pro/install.sh | bashSpool is open source and runs entirely on your machine. Star us on GitHub if this resonates.