Self-hosted personal AI agent framework in Python with a browser WebUI, terminal client, chat app channels, tools, memory and MCP support.
FoundationAgents/OpenManus
An open-source general AI agent framework that runs LLM-driven agents from a terminal, with tool use, MCP support and browser automation.
Open-source alternative to:Manus
- Stars
- 58,743
- Forks
- 10,187
- Open issues
- 441
- Last commit
- 20 days ago
What is it?
OpenManus is a Python framework for building general-purpose AI agents that take a task in natural language and carry it out with tools. You run it from the terminal, type in an idea, and the agent plans and executes steps. It is positioned as an open alternative to closed agent products that require invite codes.
The project connects to LLM APIs through a TOML configuration file, where you set the model, base URL and API key, plus an optional separate model for vision. It ships with several entry points: a general agent via main.py, an MCP tool version via run_mcp.py, and an experimental multi-agent flow via run_flow.py. A DataAnalysis agent for data analysis and visualization can be enabled in the flow configuration.
For web tasks, OpenManus uses Browser Use CLI as its default MCP server, which attaches to local Chrome or Chromium or to a Browser Use Cloud browser. BrowserGym is also supported and needs a Playwright browser installed.
Key features
- 01Runs general AI agents from a terminal using a natural-language task prompt.
- 02Configures LLM providers and a separate vision model through a TOML file.
- 03Provides an MCP tool version and an experimental multi-agent flow.
- 04Includes a DataAnalysis agent for data analysis and visualization.
- 05Automates browsers via Browser Use CLI, with local Chrome or cloud browser options.
Use cases
- Prototyping a general-purpose agent that answers questions and completes multi-step tasks.
- Experimenting with tool-using agents that browse the web on your behalf.
- Running data analysis and chart generation tasks through an LLM-driven agent.
- Studying an open agent codebase built around LLM APIs and MCP tools.
Quick start
- 1
Clone the repository
git clone https://github.com/FoundationAgents/OpenManus.git - 2
Install dependencies
pip install -r requirements.txt - 3
Create your config file from the example
cp config/config.example.toml config/config.toml - 4
Run OpenManus and enter your idea in the terminal
python main.py
Things to know
- You must add your own LLM API key to config/config.toml before it can run.
- The multi-agent run_flow version is described in the README as unstable.
- Browser automation needs extra setup, such as Chromium or Playwright browsers.
Topics
FAQ
- What is OpenManus?
- An open-source general AI agent framework that runs LLM-driven agents from a terminal, with tool use, MCP support and browser automation. OpenManus is a Python framework for building general-purpose AI agents that take a task in natural language and carry it out with tools. You run it from the terminal, type in an idea, and the agent plans and executes steps. It is positioned as an open alternative to closed agent products that require invite codes.
- Is OpenManus free to use? What license does it have?
- OpenManus is open source under the MIT license. Check the license terms before using it in commercial products.
- Is OpenManus still maintained?
- Actively maintained. The last commit to OpenManus was on Sep 30, 2026.
- What can I use OpenManus for?
- Prototyping a general-purpose agent that answers questions and completes multi-step tasks. Experimenting with tool-using agents that browse the web on your behalf. Running data analysis and chart generation tasks through an LLM-driven agent. Studying an open agent codebase built around LLM APIs and MCP tools.
- How do I install OpenManus?
- Clone the repository: git clone https://github.com/FoundationAgents/OpenManus.git Install dependencies: pip install -r requirements.txt Create your config file from the example: cp config/config.example.toml config/config.toml Run OpenManus and enter your idea in the terminal: python main.py
- Where can I get OpenManus?
- The source code and installation instructions are on GitHub: https://github.com/FoundationAgents/OpenManus
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
An open agent framework where the AI agent gets its own Dockerized Linux desktop, browser, and document tools to carry out real work.
Open-source long-horizon super agent harness that orchestrates sub-agents, memory, and sandboxes to research, code, and create across tasks lasting minutes to hours.
Blender add-on and MCP server that let AI assistants like Claude build, modify and inspect 3D scenes in Blender through prompts.
An opinionated, batteries-included agent harness for Python that runs long multi-step tasks out of the box on LangGraph.
An open-source autonomous research agent that plans questions, gathers web and local sources, and writes cited reports with any LLM provider.