Skip to content

Latest commit

 

History

History
132 lines (94 loc) · 3.84 KB

File metadata and controls

132 lines (94 loc) · 3.84 KB

aisuite

🆕 Introducing Open Coworker — an open agent harness that runs tasks and automations on your machine. Desktop app, your own API keys, your files. Get started in 2 minutes

PyPI License: MIT Build

Simple, unified tools for building with LLMs — from a drop-in chat client to a full agent harness. Three layers, each usable on its own.


Chat Completions

A uniform client across 20+ providers using an OpenAI-compatible interface. Swap models with a string — no SDK juggling, no rewrites.

import aisuite as ai

client = ai.Client()
response = client.chat.completions.create(
    model="anthropic:claude-sonnet-4-5",
    messages=[{"role": "user", "content": "Why is the sky blue?"}],
)
print(response.choices[0].message.content)
pip install aisuite          # add provider extras, e.g. aisuite[anthropic]

examples · docs


Agent API

A lightweight Agent + Runner built on the client — tools, continuation state, and tracing, without adopting a heavy framework. The agent is the reusable definition; the runner owns execution.

import aisuite as ai

def get_weather(city: str) -> str:
    """Get the current weather for a city."""
    return f"It's sunny in {city}."

agent = ai.Agent(
    name="assistant",
    model="openai:gpt-5.5",
    instructions="Answer briefly. Use tools when they help.",
    tools=[get_weather],
)

result = ai.Runner.run_sync(agent, "What's the weather in San Francisco?")
print(result.final_output)
pip install aisuite[agents]

examples · docs


Open Coworker

An open agent harness for automating daily tasks and more — a desktop app where an agent works in your folders, uses connectors and tools, produces artifacts, and runs scheduled automations. Bring your own API keys; your files and keys stay on your machine.

  • Multi-folder file access with per-folder read-only / read-write grants
  • Connectors & tools — browser automation, integrations, MCP servers
  • Artifacts — Markdown, images, PDF, CSV, spreadsheets, and Office files
  • Automations — scheduled runs that continue as conversations
  • Your models — OpenAI, Anthropic, Gemini, and local models via Ollama

Download for macOS / Windows · quickstart · docs


How they stack

Each layer builds on the one below it, and each is usable on its own:

Open Coworker        an agent harness for tasks & automations  (app)
      │  built on
Agent API            Agent + Runner: tools, state, tracing      (lib)
      │  built on
Chat Completions     one client, many providers                 (lib)

Start at whatever layer fits: call a model, wrap it in an agent, or run the whole harness.


Repository layout

libs/
  aisuite-py/     the `aisuite` package — Chat Completions + Agent API
  aisuite-js/     JavaScript/TypeScript port
apps/
  opencoworker/   the Open Coworker harness (desktop app + server)
  code-cli/       aisuite-code, the command-line coding agent
examples/         runnable examples per product
docs/             chat / agents / coworker

Documentation

Contributing

We welcome contributions! See CONTRIBUTING.md to get started.

License

MIT