Vibetracking

Track and visualize your AI coding tool usage

Website • Quick Start • Methodology • Contributing

---

What is Vibetracking?

Vibetracking aggregates usage statistics from AI coding assistants and presents them in a gamified interface. Compare your AI coding habits with developers worldwide.

• Global Leaderboards - Ranked by estimated API spending
• Personal Dashboards - Activity heatmaps and usage patterns
• Model Breakdowns - See which AI models you use most
• Fun Comparisons - Your tokens = novels written, marathons run
• Streak Tracking - Track your coding consistency

Supported Tools

Tool	Data Location
Claude Code	~/.claude/projects/*/conversation.jsonl
Cursor	CSV from cursor.com API
Codex	~/.codex/sessions/*/conversation.jsonl
Gemini	~/.gemini/tmp/*/chats/session-*.json
OpenCode	~/.local/share/opencode/storage/message/**/*.json
Amp	~/.ampcode/sessions/**/*.json
Droid	~/Library/.../googleAiStudio/history/*.json

---

Quick Start

# Scan your local AI tool data and open your dashboard
bunx vibetracking

This will:

1. Scan your machine for AI coding tool session files
2. Fetch pricing data for accurate cost estimates
3. Open your browser to import the data via GitHub OAuth

---

How We Count

Token Types

We track 5 distinct token types for accurate cost calculation:

Type	Description
Input Tokens	Tokens sent to the model (prompts, context, files)
Output Tokens	Tokens generated by the model (responses, code)
Cache Read	Cached prompt tokens reused from previous requests
Cache Write	Prompt tokens cached for future requests
Reasoning	Thinking/chain-of-thought tokens (extended thinking models)

Cost Calculation

Formula:

cost = (input × input_rate)
     + ((output + reasoning) × output_rate)
     + (cache_read × cache_read_rate)
     + (cache_write × cache_write_rate)

Pricing Sources:

• LiteLLM - Primary source for most models
• OpenRouter - Fallback for specialized models

Pricing data is cached for 24 hours.

Data Parsing

• Claude Code - JSONL with global deduplication via messageId:requestId composite keys
• Codex - JSONL with delta calculation from total_token_usage
• Cursor - CSV from cursor.com API, cache write calculated from input differentials
• Gemini - JSON with tokens.thoughts mapped to reasoning tokens
• OpenCode, Amp, Droid - Standard JSON parsing with direct token extraction

Privacy

• Local scanning only - All parsing happens on your machine via the CLI
• Explicit submission - Data only leaves your device when you click submit
• No telemetry - We don't track your usage or send analytics
• Open source - The CLI is fully open source

---

Project Structure

vibetracking/
├── src/                    # Next.js web application
│   ├── app/               # App Router pages & API routes
│   ├── components/        # React components
│   └── lib/               # Utilities & Supabase client
├── packages/
│   ├── cli/               # Bun-based CLI tool
│   └── core/              # Native Rust module (NAPI-RS)
├── supabase/migrations/   # Database schema
└── public/                # Static assets

Tech Stack

Component	Technology
Web Framework	Next.js 16 + React 19
Database	Supabase (PostgreSQL)
Authentication	GitHub OAuth
Styling	Tailwind CSS v4
Charts	Recharts v3
CLI Runtime	Bun
Native Parsing	Rust (NAPI-RS)
Deployment	Vercel

---

Development Setup

Prerequisites

• Bun >= 1.0
• Node.js >= 22
• Rust (for native module)
• Supabase account
• GitHub OAuth app

Install & Run

# Clone and install
git clone https://github.com/lfglabs-dev/vibetracking.dev.git
cd vibetracking.dev
bun install

# Build native module
cd packages/core && bun run build && cd ../..

# Start dev server
bun dev

Environment Variables

Create .env.vibetracking:

NEXT_PUBLIC_SUPABASE_URL=https://your-project.supabase.co
NEXT_PUBLIC_SUPABASE_ANON_KEY=your-anon-key
SUPABASE_SERVICE_ROLE_KEY=your-service-role-key

---

Data Flow

┌─────────────────────────────────┐
│        Your Machine             │
│  Claude, Cursor, Codex, etc.    │
└───────────────┬─────────────────┘
                │
                ▼
┌─────────────────────────────────┐
│   CLI (bunx vibetracking)       │
│   - Parallel file scanning      │
│   - SIMD-accelerated parsing    │
│   - Price lookup & aggregation  │
└───────────────┬─────────────────┘
                │
                ▼
┌─────────────────────────────────┐
│   Browser Import                │
│   /import#encoded_data          │
│   GitHub OAuth login            │
└───────────────┬─────────────────┘
                │
                ▼
┌─────────────────────────────────┐
│   Supabase (PostgreSQL)         │
│   users, daily_activity,        │
│   token_usage, user_stats       │
└─────────────────────────────────┘

---

Contributing

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

---

Credits

The CLI is based on tokscale by @junhoyeo.

---

License

MIT