I built a CLI tool for fast full-text search across codebases, designed to power file suggestions in AI coding assistants like Claude Code.
File suggestion in coding assistants needs to be fast - users expect results as they type. Traditional grep/ripgrep are great for one-off searches, but re-scanning thousands of files on every keystroke isn't practical.
GitHub:
Feedback welcome! Especially interested in ideas for detecting deleted files during incremental indexing without a full scan.
