token-bench
MITMeasure what it costs an AI coding agent to work in your codebase. An agent reads a whole file to change any part of it, so token-bench reports that per-file edit cost in Claude's own tokens and flags which files are worth splitting.
100
Security score
0 known advisories in v0.1.4
Weekly downloads
23
Unpacked size
105.8 kB
Dependencies
4
Last publish
26 days ago
Security advisories
No known vulnerabilities affect v0.1.4.
Downloads — last 30 days
679 total▼ 45.2% vs prior week
Jul 21Aug 19