Palantyra SDK

AI observability platform for LLM applications with automatic tracing.

Introduction

Palantyra is a powerful AI observability platform designed for LLM applications. The Python SDK provides automatic tracing, cost tracking, and performance monitoring for your AI-powered applications with zero-code integration.

Key Features

• ✨ Zero-code tracing - Just initialize and go
• 💰 Automatic cost tracking - Real-time cost calculation
• 📊 Performance monitoring - Latency, tokens, error rates
• 🔗 Session correlation - Group related LLM calls
• 🌐 Multi-provider support - OpenAI, Anthropic, Google
• ⚡ Async support - Works with httpx, aiohttp

Quick Installation

pip install palantyra[full]

Simple Example

import palantyra
from openai import OpenAI

# Initialize once
palantyra.initialize(api_key="your-api-key")

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