Real-time context-window monitoring for Claude Code. See your token usage, catch degradation early, and understand your AI spend — without leaving the terminal.
From the live statusline in front of you to multi-week cost reports — each level answers a different question about how you use Claude Code.
Context zone, Model Intelligence score, and token delta on every prompt — so you act now, not after the context is already gone.
context-stats graph
Export a full session to Markdown: cache efficiency, interaction timeline, zone history, and the exact moments context compacted.
context-stats export
Cost breakdown, cache analysis, and cross-project patterns over weeks or months — so you know exactly where your tokens go.
context-stats report
Pure Python 3.9+, zero external dependencies. Copy a command, wire the statusline, restart Claude Code.
Pick pip or uv — both install the same package from PyPI.
$ pip install context-stats
$ uv pip install context-stats
Add this to ~/.claude/settings.json, then restart Claude Code.
{"statusLine": {"type": "command", "command": "claude-statusline"}}
Need the full walkthrough? See Install in 60 seconds.
Free. MIT licensed. Zero external dependencies. All your data stays local — always.