Preset: Deploy Reachy Voice Robot {#default}

Deploy the full voice conversation stack on your Jetson device. The robot will listen, think with a local AI model, speak, and express emotions — all under 1 second latency.

What gets deployed:

• Robot Control — motor, camera, and sensor management
• Conversation Engine — AI dialogue + emotion system + web dashboard
• Vision Analysis — face detection, emotion recognition, and person tracking (GPU-accelerated)
• Local AI Model — powers the robot's thinking ability (auto-installed if not present)

Prerequisites:

• Reachy Mini connected to Jetson via USB
• Jetson with JetPack 6.x, SSH access, and internet

Step 1: Deploy Speech Service {#speech_service type=docker_deploy required=true config=devices/speech.yaml}

Deploy the GPU-accelerated speech recognition (ASR) and voice synthesis (TTS) service. The pre-built image includes all dependencies and models — just pull and run.

Target: Remote Deployment {#speech_remote type=remote config=devices/speech.yaml default=true}

Deploy to your Jetson over SSH with one click.

Wiring

1. Connect your Jetson to the network
2. Enter the Jetson's IP address and SSH credentials
3. Click Deploy — the system will pull the pre-built image and start the service automatically

Deployment Complete

Speech service is running at http://<jetson-ip>:8621. Quick test:

# Check service health
curl http://<jetson-ip>:8621/health
# Expected: {"asr": true, "tts": true, "streaming_asr": true}

Troubleshooting

Target: Local Deployment {#speech_local type=local config=devices/speech_local.yaml}

Deploy directly on the current machine (requires NVIDIA GPU).

Wiring

1. Ensure Docker and NVIDIA Container Toolkit are installed
2. Click Deploy to start installation

Note: First startup may take 10-15 minutes for Docker image download and model initialization.

Deployment Complete

Speech service is running at http://localhost:8621. Quick test:

# Check service health
curl http://localhost:8621/health
# Expected: {"asr": true, "tts": true, "streaming_asr": true}

Troubleshooting

Step 2: Deploy Reachy Voice Robot {#reachy_deploy type=docker_deploy required=true config=devices/reachy.yaml}

Deploy the robot control, conversation, and vision services to your Jetson. The system will automatically install Ollama and pull the AI model if not present.

Target: Remote Deployment {#reachy_remote type=remote config=devices/reachy.yaml default=true}

Deploy to your Jetson over SSH with one click.

Wiring

1. Connect Reachy Mini to Jetson via USB cable
2. Ensure the Jetson is on the network and SSH is accessible
3. Enter the Jetson's IP address and SSH credentials
4. Configure the data directory (default: ~/reachy-data) for captures and face database
5. Choose the Vision Backend:
   • Local (default) — runs vision-trt on this Jetson. Needs the USB camera attached and builds TensorRT engines on first boot (~3-5 min).
   • Cloud — skips the local vision-trt container and points the robot at a remote vision service. Fill in the Vision Service URL (e.g. tcp://192.168.1.50:8631) when selecting this option.
6. Optionally enable Kiosk Mode to auto-launch the dashboard fullscreen on boot
7. Click Deploy — the system will:
   • Install Ollama and pull the AI model if not present (~1.5 GB)
   • Pull and start robot control, conversation, and (for Local backend) vision services

Deployment Complete

The robot should start talking within 30 seconds after deployment. Open the dashboard to monitor activity:

http://<jetson-ip>:8640

Default mode: Monologue — the robot automatically generates "inner thoughts" every 5 seconds. No user interaction needed.

To check all services are running:

ssh user@<jetson-ip> "docker ps --format 'table {{.Names}}\t{{.Status}}'"

Troubleshooting

Target: Local Deployment {#reachy_local type=local config=devices/reachy_local.yaml}

Deploy directly on the current machine (requires NVIDIA Jetson with Reachy Mini connected).

Wiring

1. Connect Reachy Mini to the machine via USB cable
2. Ensure Docker and NVIDIA Container Toolkit are installed
3. Click Deploy to start installation

Note: First startup may take 5-10 minutes for Docker image download and model initialization.

Deployment Complete

The robot should start talking within 30 seconds after deployment. Open the dashboard to monitor activity:

http://localhost:8640

Troubleshooting

Deployment Complete

Your Reachy Mini voice robot is now running!

What's Happening

The robot is in Monologue Mode — it automatically generates thoughts and speaks them out loud every few seconds, with matching head movements and antenna expressions. This is ideal for exhibitions and demos.

Service Overview

Next Steps

• Open the Dashboard at http://<jetson-ip>:8640 to see conversation logs and robot status
• To switch from monologue to interactive mode, edit the config on the Jetson:

  ssh user@<jetson-ip>
  nano ~/reachy-deploy/reachy-claw.jetson.yaml
  # Change: conversation.mode: conversation
  # Change: barge_in.enabled: true
  docker restart reachy-claw
  

• To change the AI model: ollama pull <model-name>, then update llm.model in the config and docker restart reachy-claw