AutoGPT is a platform for building, deploying, and managing continuous AI agents that automate complex workflows. Written primarily in Python, the project provides both a modern platform for agent creation and a classic standalone agent implementation. The repository serves as a comprehensive ecosystem for developers who want to leverage large language models and autonomous agents to solve real-world problems.
The platform consists of several interconnected components. The AutoGPT Frontend offers a user interface where developers can build custom AI agents through an intuitive, low-code block-based interface, manage workflows, deploy agents to production, select from pre-configured agents, interact with running agents, and monitor performance through analytics. The AutoGPT Server functions as the execution engine where agents run after deployment, capable of being triggered by external sources and operating continuously. The platform includes a marketplace for discovering and deploying pre-built agents, and developers can create custom blocks to extend functionality.
The repository also maintains the classic AutoGPT implementation, which includes Forge, a toolkit for building custom agent applications that handles boilerplate code and lets developers focus on agent-specific logic. The agbenchmark tool provides performance measurement for agents supporting the agent protocol standard, offering autonomous and objective evaluation in a stringent testing environment. A frontend UI component gives users a friendly interface to control and monitor agents, compatible with any agent implementing the agent protocol. A CLI tool at the repository root simplifies setup and usage of all included tools.
The project demonstrates practical agent capabilities through example implementations, including an agent that identifies trending Reddit topics and generates viral short-form videos, and another that monitors YouTube channels, transcribes videos, extracts impactful quotes, and automatically publishes summaries to social media. These examples illustrate the platform's ability to handle complex, multi-step automation workflows.
The codebase uses a dual licensing approach. Code within the autogpt_platform folder uses the Polyform Shield License, covering the in-development platform for building and managing agents. All other portions, including the classic AutoGPT agent, Forge, agbenchmark, and the classic GUI, are licensed under the MIT License. The project also publishes related work under MIT licensing in separate repositories such as GravitasML and AutoGPT Code Ability.
Self-hosting requires Docker Engine 20.10.0 or newer, Docker Compose 2.0.0 or newer, Git 2.30 or newer, Node.js 16.x or newer, npm 8.x or newer, and VSCode 1.60 or newer. Hardware requirements include a CPU with 4 or more cores, minimum 8GB RAM with 16GB recommended, and at least 10GB of free storage. The platform supports Linux, macOS 10.15 or newer, and Windows 10/11 with WSL2. A one-line setup script is available for macOS, Linux, and Windows PowerShell to automate installation and configuration.
