Unleash Your AI Ideas: OpenManus - The No-Invite-Code Agent Framework

Tired of waiting for invites to explore the world of AI agents? OpenManus lets you dive in headfirst. Built by a team from MetaGPT in just 3 hours, this straightforward framework allows you to bring your AI agent ideas to life immediately. No more waiting. Just innovation.

OpenManus: Your Gateway to Effortless AI Agent Creation

OpenManus is designed for simplicity and customizability. With this project, you gain the power to build and experiment with AI agents without the usual hurdles.

Here's what makes OpenManus a game-changer:

• No Invite Code Required: Skip the waitlist and get started instantly.
• Simple Implementation: Easy to understand and modify, even for beginners.
• Built by MetaGPT Contributors: You're benefiting from the expertise of a leading AI community.

Getting Started with OpenManus: A Step-by-Step Guide

Ready to start building? Follow these simple steps to install and configure OpenManus and begin crafting your agent.

1. Installation: Set Up Your Environment

First, create a dedicated environment to keep your project organized.

• Create a Conda Environment:

  conda create -n open_manus python=3.12
  conda activate open_manus
  

  Copy
• Clone the OpenManus Repository:

  git clone https://github.com/mannaandpoem/OpenManus.git
  cd OpenManus
  

  Copy
• Install Dependencies:

  pip install -r requirements.txt
  

  Copy

2. Configuration: Connect to Your LLM API

To bring your agent to life, you need to configure OpenManus to access a Language Learning Model (LLM) API.

• Create a Configuration File:

  cp config/config.example.toml config/config.toml
  

  Copy

• Edit Your API Keys: Open config/config.toml and add your API keys. Popular choices include the GPT-4o model via OpenAI.

  [llm]
  model = "gpt-4o"
  base_url = "https://api.openai.com/v1"
  api_key = "sk-..." # Replace with your actual API key
  max_tokens = 4096
  temperature = 0.0
  
  [llm.vision] # For agents needing visual input.
  model = "gpt-4o"
  base_url = "https://api.openai.com/v1"
  api_key = "sk-..." # Replace with your actual API key
  

  Copy

3. Quick Start: Launch Your Agent

With the setup complete, you are ready to run OpenManus and start interacting with your agent.

• Run OpenManus: Enter your idea via the terminal after running the start command.
• (For potentially unstable versions, use the alternative run command.)

Now, it's time to bring your ideas to life using this AI Agent Framework!

Contribute to OpenManus: Shape the Future of AI Agents

OpenManus is an open-source project, and contributions are encouraged! Share your expertise and help improve the platform.

• Suggestions and Contributions: Create issues or submit pull requests on GitHub.
• Contact the Maintainers: Reach out to @mannaandpoem via email at [email protected].

Roadmap: Exciting Features on the Horizon

The OpenManus team has ambitious plans for the future. Keep an eye out for these upcoming features:

• Better Planning
• Live Demos
• Replay Functionality
• RL Fine-tuned Models
• Comprehensive Benchmarks

Join the Community: Connect with Fellow Developers

Share your experiences, ask questions, and collaborate with other OpenManus users.

• Join the Feishu Networking Group: (Link provided in the original README)

Acknowledgements

The OpenManus team extends its gratitude to anthropic-computer-use and broswer-use for their foundational support. They also thank the MetaGPT community for their invaluable contributions.

Start building your AI agent with OpenManus today!