Introduction to OpenAGI

OpenAGI is an autonomous AI system designed to act on your behalf, track its own progress, and remember context using hardware-level vector embeddings. This documentation covers the core primitives involved in setting up, expanding, and monitoring your local agent.

Quick Start Guide →

From zero to running OpenAGI on your machine in 5 minutes using Python 3.11.

Memory Architecture →

How the HDC vectors, FAISS routing, and SQLite episodic layering actually work.

Prerequisites

You will need the following installed on your machine to run OpenAGI optimally:

  • Python 3.11 or higher.
  • Git for fetching remote agents via the agency-agents framework.
  • At least 8GB of RAM (16GB recommended for heavy browsing tasks).
  • A valid Groq or NVIDIA API key.

Philosophy

We believe in Local-First Automation. Your conversation history, goals, and behavioral data should never be ingested into a generalized training set without your consent. By keeping operations local, OpenAGI remains fast, private, and hyper-personalized to you.