Documentation

Everything here also lives in the repository — this page mirrors GETTING-STARTED.md so you don't have to leave the browser to get oriented.

Requirements

Go 1.25+ to build the kernel, Python 3.11+ to run skills, Node 20+ only if you want to rebuild the UI (it ships prebuilt and embedded). Any consumer machine (8 GB RAM) is enough — no account, no cloud, no Docker.

Quickstart

  1. 1 Clone the repo and build the kernel: `cd kernel && go build -o aura.exe ./cmd/aura`
  2. 2 Start it: `.\aura.exe up` — opens the control-plane UI at localhost:9080
  3. 3 Talk to it: `aura chat --graph echo "test"` needs no model at all
  4. 4 Bring a local LLM online, then run the flagship gated multi-agent example — see GETTING-STARTED.md for the full walkthrough

CLI reference (selected)

aura up Start a node and its embedded UI
aura chat REPL or one-shot conversation with a graph, streaming replies
aura do Natural-language goal → plan → gated execution
aura connect Project an existing OpenAPI spec as skills
aura publish Sign and upload a skill to a federable registry
aura why Narrate a session's root cause from its causal log
aura federate Proxy a remote node's skills into the local node

The full reference

The README covers every command, the three frozen contracts (C1 manifest, C2 graph IR, C3 channel protocol), the HTTP/WebSocket API, and the security model in full.