codeburn

name	codeburn
description	Drive CodeBurn, a free open-source local-first CLI/TUI/web/menubar tool that reads the session files already on disk from 40 AI coding tools (Claude Code, Codex, Cursor, Gemini CLI, Grok, OpenCode, and more) and breaks down token usage and dollar cost by task, model, tool, and project. Use when the user wants to see where their AI coding spend went, find and fix token waste in a Claude Code / agent setup, cap a session's budget before it runs away, compare which model is actually worth its price, check whether AI spend shipped or was reverted, or wire live usage/savings data into an agent over MCP. Triggers on: "codeburn", "npx codeburn", "AI token usage", "AI coding cost", "where did my Claude spend go", "codeburn optimize", "codeburn guard", "codeburn compare models", "codeburn yield", "token waste in CLAUDE.md", "AI spend dashboard".
allowed-tools	Bash Read Write Edit Glob Grep
compatibility	Node.js 22.13+. Run instantly with `npx codeburn` (no install needed), or `npm install -g codeburn` / `bunx codeburn` / `pnpm dlx codeburn` / `brew install codeburn` (macOS) for a permanent command. Cursor and OpenCode auto-install `better-sqlite3`. Requires at least one supported tool with session data already on disk. Everything runs locally — no wrapper, no proxy, no API keys, nothing leaves the machine. MIT license.
metadata	{"tags":"codeburn, ai-cost-tracking, token-usage, claude-code, codex, cursor, observability, cost-optimization, cli, mcp, tui, guard-hooks","platforms":"Claude, ChatGPT, Gemini, Codex","version":"1.0","source":"https://github.com/getagentseal/codeburn"}

CodeBurn — see where your AI coding spend goes

CodeBurn is a free, open-source, local-first tool that tracks AI coding token usage and cost across 40 tools and agents (Claude Code, Cursor, Codex, Gemini, Grok, OpenCode, and more), broken down by model, project, and task. It reads the session files those tools already write to disk — no wrapper, no proxy, no API keys — and prices every call from LiteLLM's daily-refreshed pricing data. Four surfaces (terminal TUI, browser dashboard, macOS menu bar / GNOME panel, and an MCP server) all read the same on-disk source of truth.

When to use this skill

• Getting a quick answer to "where did my AI budget go this week/month?" (codeburn, codeburn overview, codeburn report)
• Finding concrete token/dollar waste in a Claude Code setup — re-read files, bloated CLAUDE.md, unused MCP servers, ghost agents/skills — and applying the fix (codeburn optimize, codeburn optimize --apply)
• Capping runaway session spend before it happens, or flagging projects with known waste at session start (codeburn guard install)
• Deciding which model is actually worth its price for a given kind of work (codeburn compare)
• Checking whether AI-assisted work actually shipped, or was reverted / abandoned (codeburn yield)
• Exposing live usage/savings data to an agent mid-conversation (codeburn mcp)
• Diagnosing why a tool shows $0 or no data (codeburn doctor)

When not to use this skill

• The user wants to change how much a model costs per call (pricing policy, rate limits) — CodeBurn only measures and reports spend, it does not set prices or throttle API calls itself (guard's caps stop a session, they don't change billing)
• The user wants cloud/team-wide spend analytics with a hosted backend as the primary need — CodeBurn's sync feature is a preview push-to-your-own-URL telemetry channel, not a hosted SaaS dashboard
• The task is about the AI tools' own behavior/config unrelated to cost (e.g., writing a new Claude Code skill, debugging an MCP server's logic) and no spend/waste/token question is actually in play

Instructions

Step 1: Run it instantly, no install

npx codeburn                 # opens the interactive TUI dashboard, last 7 days

For a permanent command: npm install -g codeburn (or bunx// on macOS). Requires Node.js 22.13+.