AI that checks its work before touching your Home Assistant
Preview changes before they're written. Catch mistakes automatically.
Your token never leaves the server.
Works with Claude Code · Claude Desktop · Codex CLI · OpenCode · Gemini CLI · Hermes Agent (preview)
"When I get home, set my main lights to a warm welcome scene" — one sentence in, a fully reviewed automation out, including suggestions you might not have thought of.
What it does · Is it safe? · Get started · How it works
Early access. The core works, and this is a good time to help shape the product. Back up your configs before letting AI touch anything. Hit a problem? Open an issue.
💬 What You Can Do
Say what you want. HA NOVA figures out the rest.
| You say | What happens |
|---|
| "Check my automations for problems" | Audits against 40+ rules — catches conflicts, dead triggers, and mistakes you'd never spot manually |
| "Why didn't my motion light trigger last night?" | Replays the actual trace and tells you exactly where it failed and why |
| "Turn on the porch light at sunset and off at 11 PM" | Builds, previews, and reviews the automation before it touches your config |
| "What happened to my living room temperature overnight?" | Pulls history and summarizes the timeline in plain language |
| "Add a weather card to my main dashboard" | Reads the dashboard, shows the change, writes only after you confirm |
| "Move all kitchen devices to the Kitchen area" | Reassigns entities and devices — shows what will change first |
| "Turn off the upstairs lights" | Turns them off, confirms the new state |
| "Show me all temperature sensors" | Finds entities by name, room, area, or label |
Automations and scripts get the full workflow today: preview, review, and a 40+ rule audit. Helpers, dashboards, and device control get preview-and-verify — deeper audits are rolling out with every release.
🛡️ Safe by Design
We built this because we didn't trust AI with our own config either.
What happens when AI wants to change something:
- Researches first. Finds your devices, checks your setup, resolves entity names. No guessing.
- Shows you the change. Full diff, before anything is written.
- You approve it. Deletes require a specific confirmation code — not just "yes."
- Writes and verifies. Reads the config back to confirm the change stuck.
- Audits itself. Checks for mistakes, conflicts, and reliability issues.
- Lets you take it back. Reply
revert to undo the last update — or use Home Assistant Backups anytime.
The ground rules — always:
- 🔒 Your HA token stays on your server. The AI never sees it.
- 🔑 Credentials live in your OS keychain, not config files.
- 📖 Every rule the AI follows is a markdown file you can read.
- 🏠 Nothing leaves your network. No cloud relay, no telemetry.
🚀 Get Started
One installer. One wizard. Done. It even finds your Home Assistant automatically.
You need: Home Assistant OS or Supervised.
- Open the latest release
- Copy the one-liner for your OS
- Run it — the wizard discovers your HA, sets up the connection, and configures your AI client
Once it finishes, try: "Show me all my automations."
Platform notes
macOS — fully tested on real hardware.
Linux — same installer, passes automated checks. Real-machine testing is ongoing.
Windows — ships x64 builds (ARM64 works via x64 emulation).
- Claude Code is the most tested client on Windows today
- Gemini CLI has basic validation
- Codex and OpenCode are still early
- Native prerequisites: Claude Code needs Git for Windows / Git Bash; Gemini CLI needs Node.js
- See .claude/INSTALL.md and .gemini/INSTALL.md
