対話型デスクトップロボット — デプロイガイド

プリセット

Jetson All-in-One

Speech recognition and synthesis, the local LLM, vision and robot control all run on one Jetson.

  • Robot: Reachy Mini connected to the Jetson via USB.
  • Network: The Jetson is reachable over SSH and has internet access during deployment to pull images.
1

Deploy Speech Service

Deploys the speech recognition (ASR) and voice synthesis (TTS) service; the image includes the models.

デプロイ方法

Deploy over SSH to a Jetson Orin NX 16GB running JetPack 6.x.

配線
  1. Connect the Jetson to the network
  2. Enter the Jetson's IP address and SSH credentials
  3. Click Deploy
入力が必要な項目(SenseCraft Solution アプリで入力)
  • Device IP*
  • SSH Username*既定値 recomputer
  • SSH Password*
既定値のある残り 1 項目
  • Speech Profile (ASR + TTS)既定値 jetson-edgellm-v090-moss
    1 件の選択肢
    • · Qwen3 ASR + MOSS-TTS-Nano (Yuewen voice)

    ASR + TTS engine combination on the v0.9.0 voice stack. Qwen3 ASR is multilingual (Chinese + English); MOSS-TTS-Nano gives the most natural, expressive voice (default voice: Yuewen).

トラブルシューティング
SymptomFix
SSH connection failedTry ssh username@ip from your computer first and check the IP and credentials
Image pull slowThe image is ~8 GB compressed; ensure stable internet on the Jetson
Service not startingRun ssh user@ip "cd reachy-jetson-voice && docker compose logs"
Health check failsFirst startup takes ~40 seconds for model warmup; wait and retry
デプロイ後

Open http://<jetson-ip>:8621/health; it returns {"asr": true, "tts": true, "streaming_asr": true}.

2

Deploy Edge LLM Chat Service

Deploys the Qwen3.5-4B chat service on the same Jetson. First startup takes about 10 minutes to download ~3 GB of model files and warm up, and the service uses about 6 GB of GPU memory.

デプロイ方法

Deploy over SSH to the Jetson from Step 1; the SSH credentials carry over.

配線
  1. Click Deploy
入力が必要な項目(SenseCraft Solution アプリで入力)
  • Device IP*
  • SSH Username*既定値 recomputer
  • SSH Password*
トラブルシューティング
SymptomFix
Health check times outRun docker logs -f edge-llm-chat-service to watch the download
Out of memory during warmupClose other GPU workloads and redeploy
/v1/models returns 502Wait until the logs print Uvicorn running, then retry
デプロイ後

Open http://<jetson-ip>:11435/v1/models; the model list includes Qwen/Qwen3-4B-AWQ.

3

Deploy Reachy Voice Robot

Deploys the robot control, conversation and vision services. The Edge LLM service from Step 2 must already be running.

デプロイ後

The robot is ready about 30 seconds after deployment.

What's Happening

The robot starts in conversation mode: talk to it and it replies in one sentence with a matching emotion and head/antenna motion.

Service Overview

The dashboard is on port 8042; other ports: robot control 38001, vision 8630, Edge LLM 11435, speech 8621.

Next Steps

  • Open the dashboard at http://<jetson-ip>:8042 to see conversation logs and robot status and to adjust settings.
  • To change the persona, edit the active speech profile's instructions.txt. To tune mic gain audio_volume, VAD sensitivity client_vad_threshold or tts_speed, edit ~/reachy-jetson-llm/reachy-voice.yaml.
  • After editing these files or changing settings on the dashboard, run docker restart reachy-voice.
デプロイ方法

Deploy over SSH to the Jetson from Step 1.

配線
  1. Connect Reachy Mini to the Jetson via USB cable
  2. Enter the Jetson's IP address and SSH credentials
  3. Configure the data directory (default ~/reachy-data) for captures and the face database
  4. Optionally enable Kiosk Mode to open the dashboard fullscreen on boot
  5. Click Deploy
入力が必要な項目(SenseCraft Solution アプリで入力)
  • Jetson IP*
  • SSH Username*既定値 recomputer
  • SSH Password*
既定値のある残り 3 項目
  • Data Directory既定値 ~/reachy-data

    Directory for persistent data (captures, face DB, etc.). Example: /mnt/usb/reachy-data

  • Voice Assistant Host既定値 127.0.0.1

    IP address of the voice assistant (speech service + LLM). Default 127.0.0.1 for local deployment.

  • Enable Kiosk Mode既定値 false
    2 件の選択肢
    • · Yes — auto-launch fullscreen dashboard on boot
    • · No — manual access only

    Auto-launch dashboard in fullscreen on device boot (for exhibition/kiosk use)

トラブルシューティング
SymptomFix
Slow reply (>10 s)Run docker logs edge-llm-chat-service and open http://<jetson-ip>:11435/v1/models to check the Edge LLM
Robot not movingReplug the USB cable and run docker restart reachy-daemon
No audio outputCheck Reachy Mini's built-in speaker and audio.device in the config
Dashboard not loadingWait 30 seconds, then open http://<jetson-ip>:8042/health
No camera feedThe vision service builds its engines on first boot (~5 min); run docker logs vision-trt
Camera not found on bootThe vision service retries automatically for about 90 seconds; wait
Camera drops after hoursPhysically replug the Reachy USB cable
デプロイ後

Open http://<jetson-ip>:8042 and say something to the robot; it replies.