SuperAGI
An open-source AI agent framework that runs locally in Docker, supports local LLMs with a GPU, and provides a browser interface for managing agents.
SuperAGI is a self-hosted AI agent framework for developers who want to build task automation and manage agents through a browser interface. It runs locally in Docker and can use local LLMs with a GPU. The Python project uses the MIT license, so you can modify it for your own applications.
Agents can run concurrently. Toolkits connect them to external systems and third-party plugins, while a toolkit marketplace supplies extensions for agent workflows. An action console lets you give agents input and permissions during a run, which matters when automated work still needs human decisions.
Agent memory storage retains information for later runs, and connections to multiple vector databases support workflows that draw on stored knowledge. ReAct workflows give agents predefined steps for automated tasks. The framework also supports custom fine-tuned models for business-specific uses, performance telemetry for inspecting agent behavior, and controls for token usage.
The repository documents local Docker and DigitalOcean deployment. Local deployment puts the agent framework on your hardware, while toolkits can still interact with external services. Its browser interface opens on the local Docker deployment.
