fff at 7.2K Stars: Rust-Powered SIMD File Search Toolkit Gives AI Agents Faster Search Than Built-In Tools via MCP

GitHub·medium signal

dmtrKovalenko/fff is a file search toolkit using Smith-Waterman fuzzy scoring with SIMD-accelerated matching, claiming to be faster than ripgrep and fzf for any long-running process. Its MCP server gives coding agents (Claude Code, Codex, Cursor, Cline) a file search tool that is faster and more token-efficient than built-in search. Features include typo-resistant path/content search, frecency-ranked access, background watching, and multi-pattern OR search using SIMD Aho-Corasick.