Spool builds search engine for your data

The pitch is strong because it targets a real pain point: AI work produces scattered context, and Spool tries to make that context queryable without shipping it to the cloud.

• –Watches Claude and Codex session directories in real time, so recent agent work becomes searchable immediately
• –Pulls in bookmarks and stars from 50+ platforms, which makes it more of a unified personal retrieval layer than a single-purpose app
• –Keeps everything local with SQLite + FTS5, which is a pragmatic fit for privacy-sensitive developer workflows
• –The macOS/Apple Silicon-only install and early-stage positioning make it feel promising but still narrow in reach
• –The `/spool` Claude Code skill hints at a useful agent loop: search your history, then feed relevant fragments back into the current task