AutoGPT

Getting Started

1. Install prerequisites — Ensure you have Docker Engine 20.10+, Docker Compose 2.0+, Git, Node.js 16+, and npm 8+ installed on your system.
2. Run the setup script — On macOS/Linux, run curl -fsSL https://setup.agpt.co/install.sh -o install.sh && bash install.sh to set up the platform automatically.
3. Configure your environment — Add your LLM API keys and adjust settings via the provided configuration files or the platform UI.
4. Build and deploy agents — Use the visual interface to compose agent workflows and deploy them to run autonomously on your local or cloud infrastructure.

Key Features

• Autonomous agents — Agents can break down goals into subtasks and execute them independently without step-by-step human input.
• Latest release (late February 2026) — The most recent release brings continued improvements to agent reliability, platform stability, and self-hosting setup.
• Visual workflow builder — Compose multi-step agent pipelines through a graphical interface rather than writing custom orchestration code.
• Latest release (March 2026) — The most recent release continues to improve agent reliability, platform stability, and the visual workflow editor.
• Self-hostable — Run the entire platform on your own hardware using Docker Compose, keeping data and API keys under your control.
• Multi-LLM support — Connect agents to various LLM backends, not limited to a single provider.
• Cloud-hosted option — A managed cloud version is available via waitlist for teams who prefer not to self-host.
• Latest release (mid-March 2026) — The most recent release continues to improve agent reliability, workflow stability, and platform performance.
• Active community — Large open-source community with thousands of members on Discord and ongoing contributions on GitHub.
• Latest release (March 2026) — The most recent release brings continued improvements to agent reliability, workflow stability, and platform performance.