contextburn
ai.arsentevPublished by arsentev.aiv0.2.1Updated Oct 24, 2026
Run efficiency for coding agents: share of paid tokens that became output, not context re-reading.
Overview
Lets a coding agent read its own token run efficiency from local Claude Code transcripts, showing how much paid context became output.
- What you can do
- The package ships a dependency-free stdio MCP server with two tools: run_efficiency returns the token and cost-weighted shares as structured data, and spend_breakdown returns the full report. It parses the Claude Code transcripts already on the machine, deduplicates usage records by message id, and weights the cost share with per-model prices kept in the script. Nothing leaves the machine.
- When to use it
- Useful when you want an assistant to check mid-session how much of the paid context is being re-read rather than turned into output, or to compare run efficiency across sessions, models and working styles. It is aimed at Claude Code users who already have local transcripts.
- Requirements
- Runs locally over stdio, launched for example with uvx from the PyPI package contextburn. It needs Python 3 and reads local Claude Code transcripts under ~/.claude/projects. No accounts, API keys or network access are declared. Optional settings include CONTEXTBURN_LANG, CONTEXTBURN_DAY_START, CONTEXTBURN_WARN and CONTEXTBURN_ALARM.
Installation
In SourceWeft
- Open contextburn in the dashboard and add it to a workspace.
- Enable the server for the chats that should use its tools.
Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.
Other MCP clients
Follow the launch instructions in the repository.
README
contextburn reads the transcripts Claude Code already writes on your machine and tells you what share of the tokens you paid for became model output — and how much was the agent re-reading context it had already sent.
Token counters answer "how much did I spend?". This answers "how much of it was work?" — a normalised share, so it can be compared across sessions, models and ways of working.
Try it
Demo
Real output over the session logs of the 36 runs behind the U-curve report — nothing else on the machine. Video with DOI: 10.5281/zenodo.22713920. The runs themselves are open: Hugging Face (DOI 10.57967/hf/10366) · Kaggle · OSF (DOI 10.17605/OSF.IO/5QTWY).
Why two numbers
- By tokens the share barely moves. Every agent step resends the accumulated context, so re-reading dominates whatever you do — it describes the agent.
- Cost-weighted the share does move, because cached reads are priced far below fresh input and output. It depends on how you run sessions — it describes you.
The comparison above comes from a controlled experiment with its dataset and analysis scripts: Clear Every Third Task: A Measured U-Curve in the Context Economy of Coding Agents.
How it counts
- Reads local Claude Code transcripts (
~/.claude/projects/**/*.jsonl). Nothing leaves the machine — no network calls at all. - Deduplicates usage records by message id and keeps the element-wise maximum. A streaming runtime writes an early snapshot and a final record for the same call: counting both double-counts it, keeping only the first halves the output.
- Weights the cost share with per-model prices kept at the top of
bin/contextburn. Update them there when they change.
Commands
Configuration
The language file exists because the menu-bar app is launched from Finder, where environment
variables never reach it: echo ru > ~/.config/contextburn/lang switches both the app and the CLI.
MCP server
Let the agent read its own run efficiency mid-session. The package ships a dependency-free MCP
server (stdio) with two tools: run_efficiency returns the shares as structured data, and
spend_breakdown returns the full report.
Or install it as a Claude Code plugin, which registers the same server:
Editor extensions
- VS Code-compatible editors (VSCodium, Cursor, Windsurf, Gitpod…) — Open VSX: arsentev-ai.contextburn. A status bar meter over the local CLI; source in
editors/vscode. - Raycast — source in
editors/raycast, Store submission pending.
Menu-bar app (macOS)
app/main.swift is a small status-bar app. It polls contextburn --json once a minute and shows the
current burn rate with an hourly graph; click a bar to see that hour's breakdown.
Set CONTEXTBURN_BIN=/path/to/contextburn if the CLI is not in ~/bin or the usual Homebrew paths.
Limits
- Claude Code transcripts only, for now.
- The cost-weighted share is only as current as the price table in
bin/contextburn.
Citing
Software DOI (all versions): 10.5281/zenodo.22712985. GitHub's "Cite this repository" button gives the
reference; metadata is in CITATION.cff.
Author
Evgenii Arsentev — arsentev.ai · ORCID 0000-0002-9120-7298
This project was published as tokmon on its first day and renamed to avoid confusion with
unrelated tools of that name; TOKMON_* environment variables still work.
License
MIT — see LICENSE.
Source: README.md at commit 2f7ac04
Tools
0Version history
1- v0.2.1LatestSep 16, 2026
Score
50% of the applicable checks met
How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.
- Reached in the last 30 days25 pts
Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.
- Auth declared10 pts
Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.
- Tools listed15 pts
Answer tools/list without a credential, or ship a package that starts in a sandbox without one.
- DescriptionRequired15 pts
- Logo10 pts
- Documentation15 pts
- License10 pts