andrewyng/aisuite

Simple, unified interface to multiple Generative AI providers

What it solves

aisuite is a lightweight Python library designed to simplify building with Large Language Models (LLMs) by removing the friction of dealing with multiple different provider SDKs. It provides a unified interface for interacting with various AI models and a higher-level framework for creating agents that can use tools.

How it works

The library is structured in two primary layers:

  1. Chat Completions API: A unified, OpenAI-style interface that allows developers to switch between providers (such as OpenAI, Anthropic, Google, Mistral, and Ollama) by simply changing a model string (e.g., openai:gpt-4o to anthropic:claude-3-5-sonnet). It standardizes request and response structures across these different services.
  2. Agents API: A layer built on top of the chat API that enables the creation of agents. It allows developers to pass plain Python functions as tools, which the library automatically converts into schemas and executes. It supports multi-turn loops via a max_turns parameter and includes prebuilt toolkits for files, git, and the shell.

Additionally, it supports the Model Context Protocol (MCP) for integrating external tools without boilerplate code.

Who it’s for

Python developers building AI applications or agents who want to avoid vendor lock-in and easily experiment with different LLM providers without rewriting their integration code.

Highlights

  • Unified API: One interface for a wide array of providers including OpenAI, Anthropic, Google, Mistral, Hugging Face, AWS, Cohere, Ollama, and OpenRouter.
  • Automated Tool Calling: Converts Python functions directly into tools and manages the execution loop automatically.
  • Agentic Framework: Includes a Runner for structured work, tool policies for security (e.g., approval policies), and state stores for persisting conversations.
  • MCP Support: Native integration with the Model Context Protocol for easy tool expansion.
  • Extensible: New providers can be added via a simple adapter pattern.

Related

  • Project
    Undertone0809/promptulate

    An AI Agent application development framework that provides a concise, Pythonic SDK to build agents with planning, tool-use, and broad LLM compatibility.

  • Project
    operand/agency

    A Python library providing an Actor model framework to build scalable agent-integrated systems by connecting AI agents with software interfaces.

  • Project
    modelscope/ms-agent

    MS‑Agent is a lightweight Python framework for building autonomous LLM agents with tool‑calling, memory, multimodal input, and a skill‑retrieval system. It uses ModelScope’s Model Calling Protocol (MCP) for LLM access, provides a React/FastAPI web UI, and includes ready‑made projects for code generation, short video creation, and deep‑research workflows.

  • Project
    fetchai/innovation-lab-examples

    A collection of 80+ production-ready Python examples for building autonomous AI agents and multi-agent systems using uAgents and various LLM frameworks.

  • Project
    mozilla-ai/any-llm

    A unified Python SDK that provides a single interface to communicate with multiple LLM providers, allowing developers to switch models without changing their code.