Solves the legacy encoding problem that breaks most AI file operations. Exposes 21 tools including read_text_file, write_file, edit_file, and grep_text_files, all with automatic detection and conversion across 22 encodings: Windows-125x, ISO-8859, KOI8-R, CP866, and UTF variants with BOM handling. Built in Go, runs as a stdio transport with directory-based access control. You'll want this when working with Delphi projects, old VB6 codebases, or any PHP/HTML sites that predate UTF-8. Instead of seeing corrupted Cyrillic or Western European characters, Claude reads the actual text. Also includes tree views, regex search, line ending conversion, and encoding detection with confidence scores.
claude mcp add --transport stdio dimitar-grigorov-mcp-file-tools uvx mcp-file-tools