# CodeBurn

CodeBurn MCP server breaks down AI coding spend, tokens, and quotas from local session files. Runs during cost audits.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · ★ 11881 repo stars
- Author: getagentseal
- License: MIT · policy=link-only · description only; the file stays with the author
- Source: https://github.com/getagentseal/codeburn
- Updated: 2026-10-21
- Security check: automated, 2026-11-17 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/mcp/codeburn

## What it does

The tool reads session logs generated by Claude Code, Codex, Cursor, and other tools to calculate exact costs by project, model, and task. It processes files directly on disk without requiring an account or sign-up.

Execution requires Node.js 22.15 or newer. The dashboard displays spending tables broken down by git branches and agent activities like debugging or coding.

The system aggregates data from multiple surfaces including the terminal, desktop app, and menu bar popover using the same local session records.

## Requirements
- Needs: Node.js (commands run via npx)

## Install

### Claude Code (native platform)

**Platform CLI**
```bash
claude mcp add codeburn -- npx -y codeburn
```

### Codex

**Platform CLI**
```bash
codex mcp add codeburn -- npx -y codeburn
```

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "codeburn": {
            "command": "npx",
            "args": [
                "-y",
                "codeburn"
            ]
        }
    }
}
```

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "codeburn": {
            "type": "stdio",
            "command": "npx",
            "args": [
                "-y",
                "codeburn"
            ]
        }
    }
}
```

### Gemini CLI

**Platform CLI**
```bash
gemini mcp add codeburn npx -y codeburn
```

## When to use it
- calculate token costs per project and model
- track AI spending by git branch
- break down agent expenses by task type
- compare costs across different date ranges

## Example prompts
- Run `npx codeburn` to check AI spend in the terminal
- Add CodeBurn MCP server with `claude mcp add codeburn -- npx -y codeburn`
- Configure CodeBurn in the Cursor `mcpServers` block

## Similar
- [codeburn-claude-cost-dashboard](https://affpapa.org/en/nevibecode/skills/codeburn-claude-cost-dashboard.md) — skill · 554 installs
- [query-token-info](https://affpapa.org/en/nevibecode/skills/query-token-info.md) — skill · 11600 installs
- [surf](https://affpapa.org/en/nevibecode/skills/surf.md) — skill · 4200 installs
- [Agent Cost Report](https://affpapa.org/en/nevibecode/skills/agent-cost-report.md) — skill · 816 installs
- [cost-summary](https://affpapa.org/en/nevibecode/skills/cost-summary.md) — skill · 684 installs
- [cost-accrual-tracker](https://affpapa.org/en/nevibecode/skills/cost-accrual-tracker.md) — skill · 175 installs

From the same set: [getagentseal/codeburn](https://affpapa.org/en/nevibecode/by/getagentseal/codeburn.md) — 2 cards

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
