OpenSource-Hub
A

aisuite

SHA-256
14.2k stars·Developer Tools·SHA-256 checksum verified

aisuite is a lightweight Python library providing a unified Chat Completions API across multiple AI providers (OpenAI, Anthropic, Google, etc.), plus an Agents API with toolkits and MCP support. It also powers OpenCoworker, a desktop AI coworker.

Smart Download

Download Download Version

vapp-v0.1.1 · 44.5 MB

Unified Python library for multi-provider LLM chat and agents, plus a desktop AI coworker.

Core Features

  • Unified Chat Completions API across 10+ providers with one-line model swap
  • Agents API with multi-turn tool calling, toolkits (files, git, shell), and native MCP support
  • OpenCoworker desktop app: chat, deep research, file access, messaging, and document generation
  • Tool policies (allow/deny lists, approval) and state stores (memory, file, Postgres)
  • Minimal dependencies: install base, optionally add provider SDKs

What It Can't Do

  • You need your own API keys for each provider (or run local models via Ollama); no free tiers are included
  • OpenCoworker desktop app currently only supports macOS (Apple Silicon) and Windows 10/11 (x64); Linux installer not available
  • Tool calling accuracy depends on the underlying model's function calling capabilities; smaller models may be less reliable

Use Cases

  • Build multi-model chat applications to compare outputs from GPT-4, Claude, Gemini, etc.
  • Deploy a desktop AI assistant (OpenCoworker) for automated daily tasks like reading files, sending emails, and generating reports

Detailed Introduction

aisuite is a dual-layer Python library for LLM applications. The foundation is a unified Chat Completions API that standardizes requests and responses across over a dozen providers (OpenAI, Anthropic, Google, Mistral, Hugging Face, AWS, Cohere, Ollama, OpenRouter, etc.) – swap providers by changing a single string. On top, the Agents API provides multi-turn tool calling, ready-made toolkits (files, git, shell), native MCP (Model Context Protocol) support, tool policies, state stores, and tracing. The repo also ships OpenCoworker, a desktop AI agent that can chat, do deep research, read files, send messages, and create deliverables like PDFs and spreadsheets. Compared to LangChain, aisuite is significantly leaner (fewer dependencies, simpler API) and focuses on a single, consistent interface without complex chains or abstractions. It's ideal for developers who want to quickly swap AI providers, build custom agent workflows, or deploy a desktop AI assistant with minimal overhead.

Tags

AILLMChat CompletionsAgentsMCPMulti-ProviderOpenCoworkerPython Library

Getting Started

1

Download installer

Click the button above to download the installer for your system

2

Install the software

Open the downloaded dmg file, then drag the app to Applications

3

Install aisuite: pip install aisuite[all] or pip install aisuite[anthropic]

4

Set API keys as environment variables or in config for your target providers

5

Write your first call: import aisuite, create Client, call client.chat.completions.create(model='openai:gpt-4o', messages=...)

Install Guide
  1. Install aisuite: pip install aisuite[all] or pip install aisuite[anthropic]
  2. Set API keys as environment variables or in config for your target providers
  3. Write your first call: import aisuite, create Client, call client.chat.completions.create(model='openai:gpt-4o', messages=...)
File Integrity

SHA-256 checksum verified

Checksum extracted from GitHub official Release page

SHA256 Checksum

51696787719a46297819d93ec0c8bb98690a09b1560ac21c9074c06531a36b81

This checksum is extracted from the GitHub Release page. Verify file integrity after download.

All SHA-256 checksums on this platform are extracted from the project's official GitHub Release page, without any modification. You can independently verify them on the GitHub Releases page.

Open Source Transparency

View GitHub Source
Environment Guide

Uninstall Info

Uninstall aisuite: run pip uninstall aisuite. For OpenCoworker desktop app, remove from Applications folder (macOS) or Control Panel (Windows).

No Extra Dependencies

Ready to use after download. No additional runtime required.

Project Info
LicenseMIT
Last Updated2026-06-14T06:58:16Z
GitHub RepositoryOfficial Website

Similar Projects