aisuite
A Python library that gives one simple interface for calling many AI model providers.
- Stars
- 16.4k
- Forks
- 1.7k
- Licence
- MIT
- Language
- Python
- Last update
- 2 months ago
- Type
- Library
What it is
aisuite is a Python library that lets a developer call language models from several providers, such as OpenAI, Anthropic, and Google, using one common format. You switch providers by changing a model name string like "openai:gpt-4o". It also includes an Agents layer that lets models call Python functions as tools.
- Send the same prompt to OpenAI and Anthropic models and compare the answers
- Give a model a Python function, such as a weather check, and let aisuite run the tool-call loop
- Stream model responses chunk by chunk using one loop for several providers
- Attach MCP servers, such as a filesystem server, as tools for a model
Great for
- Python developers who want to compare or switch between LLM providers without rewriting code
- Developers building simple tool-using agents with plain Python functions
- Teams that want an MCP server's tools available to a model with little setup
Not ideal for
- Non-programmers, since it is a code library and requires writing Python
- People who only want a ready-made chat app; the README points to the separate OpenWorker desktop app for that
How to get it
- 1
Install with a package manager
You'll need: Python
Install base package
pip install aisuiteInstall with all provider SDKs
pip install 'aisuite[all]' - 2
Run it
- Install the package with pip
- Set API keys for the providers you plan to use, following the Chat Completions quickstart
- Create a client with ai.Client()
- Call client.chat.completions.create with a model string such as openai:gpt-4o
- 3
Get the most out of it
- Use stream=True for incremental output across supporting providers
- Pass Python functions as tools with max_turns to let aisuite run the tool loop
- Use toolkits such as files, git, and shell for prebuilt agent tools
- Connect any MCP server with pip install 'aisuite[mcp]'
Downloads
Strengths
- One OpenAI-style interface across many providers, switched by changing one string
- Plain Python functions can be passed as tools and their schemas are generated automatically
- Has a clear split between a simple chat API and a fuller Agents API for longer tasks
Trade-offs
- You must install and configure API keys and provider SDKs yourself
- The README is truncated and says little about maturity, versioning, or stability guarantees
- Some features, such as state stores and tool policies, are described only briefly here
How it compares
| Project | Stars | Type | Level | Updated | Best for |
|---|---|---|---|---|---|
| aisuite (this) | 16.4k | Library | intermediate | 2 months ago | Python developers who want to compare or switch between LLM providers without rewriting code |
| langchain | 149k | Framework | intermediate | 4 days ago | Python developers building LLM-powered applications |
| llama_index | 52.8k | Library | intermediate | 10 days ago | Python developers building retrieval-augmented generation (RAG) apps over private documents |
Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.