ConnectOnion
Open-source Python framework for building production-ready LLM agents with plain functions as tools.
🛡️ AgentReady threat assessment
MAESTRO 7-layer threat model + OWASP AIVSS risk score for ConnectOnion, derived from its capabilities.
These scores are auto-generated from public information (the agent's own listing, docs, and repository) using the canonical OWASP AIVSS formula and the MAESTRO framework — an estimate for guidance, not a penetration test, audit, or certification. See the scoring methodology. Are you the vendor? Factual corrections are free.
Overview
ConnectOnion is an open-source Python AI agent framework for developers who want to build agents with minimal boilerplate. Its core Agent API turns regular typed Python functions into tools automatically, and the documentation shows quick setup with pip install connectonion and a two-line agent pattern. The framework supports OpenAI, Anthropic, and Gemini models, and includes documentation for CLI commands, logging, debugging, TUI components, plugins such as ReAct, Eval, Shell, approvals, skills, subagents, and built-in tool integrations such as shell, file tools, browser automation, Gmail, Outlook, Calendar, memory, and todo workflows. It is best suited for engineering teams building custom LLM agents, tool-calling assistants, multi-agent workflows, and Python-based automation without adopting a heavy framework layer.
Key features
- python
- open-source
- llm-orchestration
- tool-calling
- multi-agent
- openai
- anthropic
- gemini
- developer-tools
Use cases
- Building Python-based LLM agents with plain functions as tools
- Creating tool-calling assistants with OpenAI, Anthropic, or Gemini models
- Adding approval workflows for potentially dangerous tool execution
- Prototyping multi-agent or subagent workflows
- Building developer automation agents with shell, file, browser, memory, and calendar tools