CodeBurn
github.com/getagentseal/codeburn- Category
- Developer Tools
- Rank
- No. 251Tools index
- Pricing
- Open Source
- Type
- TOOL
- Builder
- getagentseal
- GitHub
- 10.9k stars
- Latest release
- v0.9.24
- Date
About
A local terminal dashboard that tracks AI coding token usage and costs across 18 different AI coding tools like Claude Code, Cursor, and GitHub Copilot. It analyzes spending patterns by project, model, and task type to help developers optimize their AI coding budget without requiring API keys or proxies.
What it does
CodeBurn turns existing AI coding histories into interactive views, exportable summaries, model comparisons, budget checks, and waste findings. It can also suggest configuration fixes, journal applied changes, and compare estimated savings with later activity.
Why it's ranked here
The appeal is unusually broad without abandoning local operation. One dataset feeds terminal, web, desktop, and menu bar views, while reporting extends beyond totals into task classification, model efficiency, budgets, context composition, and corrective actions.
What's good
It works from records developers already have and keeps processing local. Reports support explicit date ranges, provider filtering, plain text, JSON, and CSV. Applied configuration changes are backed up and journaled, with undo protection when files have changed afterward.
Tradeoffs
It requires Node.js 22.13 or newer for the command-line package and needs supported session data on disk. Cost accuracy depends on model pricing, aliases, overrides, and provider records. Task categories and retry counts use tool patterns and keyword heuristics, so they are useful signals rather than ground truth.
How to use it well
It suits developers or teams using several AI coding agents who want a regular spend and workflow audit. Start with the dashboard, export monthly summaries, investigate costly models or tasks, then review suggested fixes before applying them. It does not replace provider billing records or measure work whose sessions are unavailable.
Technical notes+
package.json defines an ESM npm CLI, requires Node.js 22.13+, builds with tsup, and runs tests with Vitest. src/cli.ts performs the runtime version check before dynamically importing the main module. src/main.ts wires reporting, budgets, optimization, guard, sharing, web, menu bar, export, comparison, and provider parsing commands. src/classifier.ts combines observed tool use with ordered keyword rules for task labels and retry detection, while src/bash-utils.ts conservatively distinguishes read-shaped shell activity. src/config.ts writes randomized temporary files before atomic rename, and src/codex-cache.ts fingerprints source records and persists its cache through a permission-restricted temporary file and rename.
Observed
- License
- MIT
- Primary language
- TypeScript and TSX
- Packaging
- ES module npm package with a global codeburn executable
- Install surface
- npx, npm global install, bunx, pnpm dlx, and Homebrew
- Interfaces
- Command-line interface, terminal dashboard, web dashboard, desktop application, macOS menu bar, and MCP
- Platform support
- macOS, Windows, and Linux desktop packages; Linux GNOME panel support
- Runtime requirement
- Node.js 22.13 or newer
- Test tooling
- Vitest test scripts and Playwright development dependency
Read from README.md, package.json, src/cli.ts, src/main.ts, src/budget.ts, src/config.ts, src/cli-date.ts, src/compare.tsx, src/currency.ts, src/bash-utils.ts, src/classifier.ts, src/codex-cache.ts, src/audit-report.ts, src/context-tree.ts.
What it can do
Track AI coding token usage across multiple tools
AI coding activity from 18 different tools (Claude Code, Cursor, GitHub Copilot, etc.) → Token consumption metrics and usage statistics
Calculate AI coding costs
Token usage data and pricing models → Cost breakdowns and spending totals
Analyze spending patterns by project
Project-specific AI tool usage data → Per-project cost analysis and spending reports
Break down costs by AI model type
Usage data from different AI models → Model-specific cost analysis and comparison
Categorize expenses by coding task type
AI tool usage categorized by task → Task-based spending breakdown and analysis
Monitor performance across AI coding tools
Performance metrics from multiple AI coding tools → Tool performance comparisons and efficiency reports
Generate local analytics dashboard
Aggregated AI coding usage and cost data → Terminal-based dashboard with visualized metrics
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.