Documentation

Get Started with Axon

From zero to running in under five minutes. Everything you need to deploy your self-hosted AI command center.

Quick Start

Terminal

$ git clone https://github.com/brandonkorous/axon.git

$ cd axon

$ cp .env.example .env

$ docker compose up

Clone the repository and copy the example environment file. Add your API keys to .env, then start everything with Docker Compose.

The frontend will be available at localhost:3000 and the backend API at localhost:8000.

Docker and Docker Compose installed
At least 8 GB RAM (16 GB+ recommended for local LLMs)
An API key for your preferred LLM provider (Anthropic, OpenAI) — or Ollama for fully local inference
CUDA-compatible GPU with 8 GB+ VRAM (optional, for local models)

1
Open the dashboard
Navigate to localhost:3000 in your browser.
2
Create an organization
Choose from built-in templates — Startup, Family, Job Hunt, Creator, or Student — each with curated specialist advisors.
3
Talk to Axon
Start a conversation. Axon routes your request to the right advisor and builds memory over time.
4
Enable local LLMs (optional)
Run docker compose --profile local-llm up to start Ollama for fully local inference.

Key environment variables in your .env file:

REST and WebSocket endpoints for the Axon backend.

Build custom specialist advisors with their own personas and vaults.

How neural memory works: vaults, recall, learning, and consolidation.

Model providers, org templates, voice settings, and environment variables.

The central orchestrator. Routes requests to specialist advisors based on organization type and query context.

Local LLM (Ollama) handles recall, learning, and consolidation. Reasoning model (Claude, GPT, or local) handles conversations.

Docker Compose orchestrates the backend (FastAPI), frontend (React), and optional services (Ollama, SearXNG).