LiteLLM

A simple, open-source library to call over 100 LLMs using a standardized OpenAI format, simplifying integration and management for developers.

LiteLLM is an open-source Python library designed to streamline interactions with a wide array of Large Language Models (LLMs). It provides a unified, consistent interface, allowing developers to call APIs from over 100 providers—including OpenAI, Anthropic, Google Gemini, and Cohere—using the familiar OpenAI message format. This eliminates the need to write and maintain provider-specific code, significantly simplifying the development of LLM-powered applications. Beyond simple API abstraction, LiteLLM offers powerful features for production environments. It enables developers to implement robust logic such as automatic retries, fallbacks to alternative models or deployments, and load balancing across multiple API keys. The library also includes detailed cost tracking and usage monitoring, giving teams visibility into their spending across all models. For more complex use cases, LiteLLM can be deployed as a self-hosted proxy server, acting as a central gateway for all LLM requests within an organization.