TokenMeter 🔥

Open-source, self-hosted AI Cost Intelligence Platform.

What it does

• Proxy all your LLM API calls through TokenMeter
• Track every token, every request, every rupee/dollar spent
• Dashboard real-time cost visibility per developer, per model, per feature
• Smart Routing auto-pick cheapest model that meets quality threshold
• Semantic Cache skip duplicate API calls, save 45-80%
• Waste Detection find oversized model usage, duplicates, failed requests
• Prompt Optimizer reduce token usage by removing filler/redundant instructions
• ROI Tracker cost vs quality scoring per model
• Cost Forecasting predict future spend using historical trends
• Rate Limiting per-developer token/request limits
• Prompt Library save, share, and reuse optimized prompts
• Quality Scoring auto-rate LLM response quality, model leaderboard
• Usage Heatmap visualize when API calls happen (hour/day patterns)
• Cost Tags tag requests by project/team/environment for billing
• Webhook Integration fire webhooks on budget alerts, anomalies, etc.
• Team Dashboard multi-user auth with admin/developer/viewer roles
• A/B Testing same prompt, multiple models, compare quality + cost
• Budget Management daily/monthly limits with alerts
• Slack/Discord Alerts webhook notifications
• CSV Export download reports
• Self-hosted your data stays with you

Tech Stack

• Backend: Python (FastAPI)
• Frontend: React (Vite) + Recharts
• Database: SQLite
• No cloud dependency — runs on your machine

Quick Start

# Backend
cd backend
pip install -r requirements.txt
cp .env.example .env  # add your API keys
python main.py

# Frontend
cd frontend
npm install
npm run dev

Docker

docker-compose up -d

Architecture

Your App → TokenMeter Proxy → OpenAI/Anthropic/Gemini
                ↓
         SQLite DB (logs every request)
                ↓
         React Dashboard (17 tabs, real-time stats)

API Headers

X-Developer: your-name     # track per developer
X-Feature: my-feature      # track per feature/project
X-Smart-Route: on          # auto-pick cheapest model
X-Cache: on                # enable semantic caching

Dashboard Tabs

Overview | Models | Developers | Logs | Budgets | Waste | Savings | Optimizer | ROI | Forecast | Rate Limits | Prompts | Quality | Heatmap | Tags | Webhooks | Team

Default Login

• Username: admin
• Password: admin

License

MIT — use it however you want.