coolhand 0.1.0

A Python SDK for monitoring and analyzing AI API interactions with LLM services

Coolhand Python

Automatic monitoring for LLM API calls in Python applications.

Installation

pip install coolhand

Quick Start

import coolhand  # Auto-initializes and starts monitoring

# All LLM API calls are now automatically captured
from openai import OpenAI
client = OpenAI()
response = client.chat.completions.create(
    model="gpt-4",
    messages=[{"role": "user", "content": "Hello!"}]
)

Manual Configuration

from coolhand import Coolhand

coolhand_client = Coolhand(
    api_key='your-api-key',
    silent=False,  # Enable verbose logging
)

Environment Variables

• COOLHAND_API_KEY: Your Coolhand API key
• COOLHAND_SILENT: Set to false for verbose logging

Supported LLM APIs

Automatically captures requests to:

• OpenAI (api.openai.com)
• Anthropic (api.anthropic.com)

Requirements

• Python 3.7+
• httpx (for HTTP client patching)

License

Apache-2.0