ConnectOnion is an open-source AI agent framework designed for simplicity and production readiness, targeting developers building AI agents. It enables rapid prototyping and scalable deployment with features for collaboration and debugging, allowing users to transition from simple scripts to complex production systems without significant code rewrites.

How It Works

The framework centers around a single Agent class that automatically converts Python functions into tools, adhering to a "Keep simple things simple, make complicated things possible" philosophy. It offers a streamlined API for creating agents, integrating tools, and enabling interactive debugging. Advanced features include a plugin system for composability and a trust system for multi-agent collaboration, aiming for a no-boilerplate transition from prototype to production.

Quick Start & Requirements

• Installation: pip install connectonion
• Prerequisites: Python 3.10+
• Setup: The CLI guides through API key setup automatically.
• Links: Documentation, Discord

Highlighted Details

• Simple API: Features a single Agent class with automatic function-to-tool conversion, allowing agents to be created in just a few lines of code.
• Production Ready: Battle-tested with major LLMs (GPT-5, Gemini 2.5, Claude Opus 4.1), includes built-in logging, debugging, and safety controls like max_iterations.
• Interactive Debugging: Utilizes the @xray decorator and auto_debug() function for pausing execution at tool breakpoints, inspecting state, editing variables, and running a full Python REPL.
• Plugin System: Enables packaging reusable capabilities as plugins and composing them easily across multiple agents.
• CLI Templates: The co create command offers templates (e.g., playwright, meta-agent) for rapid project setup and automatic API key configuration.

Maintenance & Community

ConnectOnion is community-driven, with over 1000 developers contributing. An active Discord community is available for support and discussion.

Licensing & Compatibility

The project is released under the MIT License, permitting commercial use and integration into closed-source projects.

Limitations & Caveats

The example calculate tool uses eval(), which requires careful security considerations for production environments. While described as "Production Ready," advanced multi-agent collaboration features may require thorough testing in specific production use cases.