Browse
→ Code & Dev
→ TokenCut
TokenCut
Local CLI and MCP server for reducing verbose AI coding output while keeping recoverable detail. It can run commands with output folding, compact pasted logs or tables, package selected files, save explicit task checkpoints, and expose a smaller coding-agent profile for symbol search, exact method reads, outlines, and guarded edit previews without language-server daemons. Public GitHub repository created September 20, 2026.
MCP unverified
Integration
| Transport | stdio |
| Auth | none |
| Endpoint | tokencut mcp --profile coding |
| Install | uv tool install git+https://github.com/00200200/tokencut.git |
Use Cases
| 01 | Run noisy tests or shell commands through an agent-facing filter that preserves diagnostics and recovery references |
| 02 | Give coding agents precise code context through symbol search, method reads, outlines, and hash-guarded edit previews |
| 03 | Reduce MCP tool schema size with a focused coding profile while keeping full CLI commands available |
Tags
token-reduction coding-agents logs command-output symbol-search python uv stdio
Machine-readable: /api/servers.json
· JSON-LD schema embedded in <head>