nabaztag-ai 🐰

AI brain for the original Nabaztag rabbit — LLM tool-use embodiment, voice pipeline, MCP server. Physical AI, 2006 edition.

(demo GIF — coming with Phase 1)

What & why

In 2006 the Nabaztag:tag was the first consumer connected device — a Wi-Fi rabbit that read the news and wiggled its ears. Twenty years later, this project revives a stock, unopened Nabaztag:tag as the body of a modern AI assistant: wake word → speech-to-text → Claude with tool use → text-to-speech, played through the rabbit's own speaker, with the LLM deciding ear positions and LED moods itself.

The rabbit stays 100% original. It talks to a self-hosted OpenJabNab server; modern audio input comes from an external reSpeaker 4-mic array; the brain runs on a UDOO Bolt — a board co-created by the project's owner, closing a personal 20-year loop between the first connected device and modern edge AI.

Full spec: docs/ARCHITECTURE.md.

Hardware BOM

Item	Role	~Price
Nabaztag:tag (V2), stock	Body: speaker, ears, LEDs, RFID, button	~€70 (second-hand)
reSpeaker Flex XVF3800 Circular-4	Ears-in: 4-mic array, beamforming, DoA	~€50
UDOO Bolt (or any Linux PC/SBC)	Brain: OJN server, rabbit-brain, MCP	you probably have one
USB Wi-Fi dongle (AP-mode capable)	Dedicated legacy WPA-TKIP AP (§4.1)	~€10

Quickstart (no hardware needed)

git clone https://github.com/<you>/nabaztag-ai && cd nabaztag-ai
python -m venv .venv && source .venv/bin/activate
pip install -e "brain[dev]"
pytest                      # runs against the mock-OJN simulator
python -m rabbit_brain.body.demo --mock-ojn   # drive a simulated rabbit

With real hardware, start from Gate S0 (the rabbit's legacy Wi-Fi segment — see docs/ARCHITECTURE.md §4.1 and ojn/).

Architecture

Nabaztag:tag ◄─Violet proto─► OpenJabNab (self-hosted) ◄─REST─► rabbit-brain
   stock, WPA-TKIP segment         on the Bolt                wake→VAD→STT→Claude(tools)→TTS
                                                                    │
reSpeaker XVF3800 ──USB audio + DoA──────────────────────► BodyController (arbiter)
                                                                    ▲
                                                     nabaztag-mcp ──┘ → Claude Desktop/Code

All body output funnels through the BodyController (single owner of the body, priority queue); bodies are swappable behind the BodyAdapter protocol.

Roadmap

• v1 — Nabaztag:tag via OpenJabNab (this repo, phases S0 → 4)
• P1 — ReachyMiniAdapter: same brain, different body (comparison protocol)
• P2 — VentunoQLocalProfile: fully local STT/TTS/LLM on Arduino VENTUNO Q

License

Apache-2.0 (root, covers brain/, mcp/, demos/, docs, configs). If Gate G0 ever requires the OpenJabNab choreography plugin, it will live in ojn/plugin_choreo/ under OpenJabNab's GPL (derivative work), cleanly separated — the brain talks to it only over HTTP.

Nabaztag is a trademark of its respective owner; this is an independent community project, not affiliated with Violet/Aldebaran.