Track how much developers burn on AI coding tools. Real usage stats, ranked by sessions, cost, and activity. Privacy-first and open source.
Be the first. Login with GitHub and run npx devburn setup
Points reflect cumulative AI coding usage across sessions, turns, and spend. The formula rewards consistency and depth.
Only aggregate stats are shared: token counts, costs, model names. Never: prompts, code, file paths, or env vars. The CLI is open source — verify exactly what gets sent.
1. Login with GitHub above.
2. Run npx devburn setup in your terminal.
3. Code normally — stats sync automatically after each session.
Auto hook: Claude Code fires a stop hook after each session, syncing stats in the background.
Manual sync: Run npx devburn sync anytime to push accumulated data.
DevBurn tracks aggregate AI coding usage: session counts, token usage, estimated costs, active days, and streaks. It never accesses your prompts, code, file paths, or environment variables.
Login with GitHub, then run "npx devburn setup" in your terminal. Your Claude Code usage syncs automatically via a stop hook. You can also run "npx devburn sync" to push stats manually.
No. DevBurn only collects aggregate statistics: token counts, cost estimates, model names, and timestamps. The CLI is open source so you can verify exactly what gets sent.
Currently Claude Code (Claude Opus, Sonnet, Haiku). Support for OpenAI Codex, Cursor, and Gemini Code Assist is planned.
Points = (sessions x 10) + turns + (cost x 100). This formula rewards consistent daily usage, deep coding sessions, and overall investment in AI-assisted development.