AutoAgent

AutoAgent lets people create AI agents by describing the work they want done. It's for users who want custom agents without writing the agent code themselves, or a ready-made research assistant for gathering information and producing reports.

• Agent and tool creation: Describe an agent's role in conversation, and the agent editor generates its profile and the tools it needs.
• Multi-agent workflows: Turn a high-level task description into a workflow of collaborating agents, with automatic workflow generation and adaptation; the workflow editor doesn't create tools.
• Research and file analysis: Use the built-in research agents to retrieve information, tackle analytical tasks, and generate reports, with file uploads as additional input.

AutoAgent is a free, open-source Python framework under the MIT license. It runs locally through a command-line interface and uses Docker for the environment where agents work.

Model support includes Claude, OpenAI models, Gemini, Mistral, DeepSeek, and OpenAI-compatible endpoints such as Grok. It also connects through Hugging Face, Groq, and OpenRouter. These API connections require provider credentials and internet access, and send model requests to the selected service. Local execution of the framework doesn't make those model calls offline. AutoAgent also requires a GitHub token; tools drawn from RapidAPI require their own subscriptions and API keys.