AI Token Monitor 🚀

Track token usage and cost for OpenRouter, Groq, and OpenAI — with FastAPI integration.

Features

• ✅ Token tracking per request and per chat session
• ✅ Cost estimation (real model pricing)
• ✅ Multi-turn chat session management
• ✅ FastAPI integration with middleware
• ✅ Production logging

Install

pip install ai-token-monitor

Quick Start

from ai_token_monitor import TokenMonitor, ChatManager
from ai_token_monitor.utils import normalize_response, extract_reply

monitor = TokenMonitor()
chat_manager = ChatManager()

chat_id = chat_manager.create_chat("user_123")
chat_manager.add_message(chat_id, "user", "Hello!")

# After getting response from OpenAI/OpenRouter:
data = normalize_response(raw_response)
answer = extract_reply(data)
monitor.track(data, model="openai/gpt-4o-mini", chat_manager=chat_manager, chat_id=chat_id)

print(monitor.summary())

Environment Variables

Copy .env.example to .env and set:

OPENROUTER_API_KEY=your_key_here

Run the FastAPI Example

uvicorn examples.fastapi_app:app --reload

Supported Models

• openai/gpt-4o, openai/gpt-4o-mini, openai/gpt-3.5-turbo
• groq/llama3-8b-8192, groq/llama3-70b-8192
• mistralai/mixtral-8x7b-instruct, mistralai/mistral-7b-instruct
• anthropic/claude-3-haiku, anthropic/claude-3.5-sonnet

License

MIT