Smart Warehouse Management

Preset: SenseCraft Cloud {#sensecraft_cloud}

Use SenseCraft cloud service for voice AI. Simplest setup - just deploy the warehouse system and connect your Watcher to SenseCraft platform.

What you'll get:

• Voice-controlled inventory management (stock in/out by speaking)
• Real-time inventory dashboard
• Works with SenseCAP Watcher out of the box

Requirements: Internet connection · SenseCraft account (free)

Step 1: Warehouse System {#warehouse type=docker_deploy required=true config=devices/recomputer.yaml}

Deploy the inventory management service with voice control and web dashboard.

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

Run the warehouse system on this computer.

Wiring

1. Ensure Docker is installed and running
2. Click Deploy button to start services

Troubleshooting

Target: Remote Deployment (R1100) {#warehouse_remote type=remote config=devices/warehouse_remote.yaml default=true}

Deploy to reComputer R1100 edge device.

Wiring

1. Connect R1100 to power and ethernet, ensure it's on the same network as your computer
2. Enter IP address reComputer-R110x.local (or check your router)
3. Enter username recomputer, password 12345678
4. Click Deploy and wait for installation to complete

Troubleshooting

Step 2: Configure Warehouse System {#warehouse_config type=manual required=true}

After deployment, open the warehouse system to complete initial setup:

1. Open browser and visit http://server-ip:2125 (use localhost for local deployment)
2. First visit will show "Set Administrator" dialog, fill in details and confirm
3. Click "Inventory List" on the left to import existing inventory (Download Excel Template)

Troubleshooting

Step 3: Configure Watcher Device {#sensecraft type=manual required=true}

Connect your Watcher to SenseCraft cloud platform:

1. Power on Watcher, connect your phone to the WiFi hotspot displayed on the device screen, configure WiFi in the popup page, device will restart and show a 6-digit verification code
2. Login to SenseCraft AI Platform, click "Watcher Agent" → "Bind Device", enter the 6-digit code shown on your Watcher
3. Click "Create" to make a new Agent, set name and language, then save
4. Click the settings icon on your Agent card, go to "MCP Setting" and copy the endpoint URL

Troubleshooting

Step 4: Connect to Agent {#mcp_bridge type=manual required=true}

This step enables Watcher to control the warehouse system:

1. Open warehouse system, go to "Agent Configuration" on the left sidebar
2. Click "Add Agent", fill in the name, paste the endpoint URL from previous step
3. Click "Save and Start"
4. Click "MCP Endpoint" on the agent card, refresh status - "Connected" means success

Troubleshooting

Step 5: Demo & Testing {#demo type=manual required=false}

Try these voice commands:

Check the warehouse web interface to see inventory changes after speaking.

Troubleshooting

Deployment Complete

Your voice-controlled warehouse system is ready!

Access points:

• Warehouse System: http://<server-ip>:2125
• SenseCraft Platform: sensecraft.seeed.cc

Try saying "Stock in 10 boxes of apples" to test voice inventory management.

Preset: Private Cloud {#private_cloud}

Self-host the voice AI server while using cloud APIs (DeepSeek, OpenAI, etc.) for LLM and TTS. Your data stays on your network - only API calls go to the cloud.

What you'll get:

• Full control over your data - inventory stays on your network
• Flexible AI model choices (DeepSeek, GPT-4, Qwen, etc.)
• Customize voice assistant prompts and behavior

Requirements: Internet connection (for LLM/TTS API) · API keys for your AI provider (recommend DeepSeek - free credits on signup)

Step 1: Warehouse System {#warehouse type=docker_deploy required=true config=devices/recomputer.yaml}

Deploy the inventory management service with voice control and web dashboard.

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

Run the warehouse system on this computer.

Wiring

1. Ensure Docker is installed and running on your machine
2. Click Deploy button to start services

Troubleshooting

Target: Remote Deployment (R1100) {#warehouse_remote type=remote config=devices/warehouse_remote.yaml default=true}

Deploy to reComputer R1100 edge device.

Wiring

1. Connect R1100 to power and ethernet, ensure it's on the same network as your computer
2. Enter IP address reComputer-R110x.local (or check your router)
3. Enter username recomputer, password 12345678
4. Click Deploy and wait for installation to complete

Troubleshooting

Step 2: Configure Warehouse System {#warehouse_config type=manual required=true}

After deployment, open the warehouse system to complete initial setup:

1. Open browser and visit http://server-ip:2125 (use localhost for local deployment)
2. First visit will show "Set Administrator" dialog, fill in details and confirm
3. Click "Inventory List" on the left to import existing inventory (Download Excel Template)

Troubleshooting

Step 3: Voice AI Service {#voice_service type=docker_deploy required=true config=devices/xiaozhi_server.yaml}

Deploy the voice AI service to enable voice interaction with Watcher. Select "Private Cloud" mode and fill in your LLM API details.

Get API keys: DeepSeek (recommended, free credits for new users) · OpenAI · Moonshot

TTS uses free Edge TTS by default, no additional configuration needed.

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

Wiring

1. Ensure Docker is installed and running
2. Click Deploy button to start services

Target: Remote Deployment (R1100) {#voice_remote type=remote config=devices/xiaozhi_remote.yaml default=true}

Wiring

1. Enter R1100 IP address and SSH credentials
2. Click Deploy and wait for installation to complete

Troubleshooting

Step 4: Configure Watcher Device {#watcher_config type=manual required=true}

Connect your Watcher to the local voice service:

1. Power on Watcher, connect your phone to the WiFi hotspot displayed on the device screen
2. In the popup page, tap the "Advanced" tab
3. Change OTA URL to: http://server-ip:18003 (server IP is the device running voice service)
4. Configure WiFi (2.4GHz only), device will restart and show a 6-digit verification code
5. Visit voice service console at http://server-ip:18003 and bind the device code

Troubleshooting

Step 5: Connect to Agent {#agent_config type=manual required=true}

Add an agent in the warehouse system to let Watcher control inventory:

1. Open warehouse system, go to "Agent Configuration" on the left sidebar
2. Click "Add Agent", fill in name, enter ws://server-ip:18004/mcp in Endpoint field
3. Click "Save and Start"
4. Click "MCP Endpoint" on the agent card, refresh status - "Connected" means success

Troubleshooting

Step 6: Demo & Testing {#demo type=manual required=false}

Try these voice commands:

Check the warehouse web interface to see inventory changes after speaking.

Troubleshooting

Deployment Complete

Your private cloud warehouse system is ready!

Access points:

• Warehouse System: http://<server-ip>:2125
• Voice Service Console: http://<server-ip>:18003

Your data stays on your network. Try saying "How many apples left?" to test.

Preset: Edge Computing {#edge_computing}

Run everything locally including LLM and TTS - no internet required. Ideal for air-gapped environments or strict data compliance.

What you'll get:

• 100% offline operation - works without internet
• All data stays within your local network
• Local LLM inference at ~16 tokens/sec

Requirements: reComputer J4012 or equivalent Jetson device · Internet needed for initial deployment only

Step 1: Warehouse System {#warehouse type=docker_deploy required=true config=devices/recomputer.yaml}

Deploy the inventory management service with voice control and web dashboard.

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

Run the warehouse system on this computer.

Wiring

1. Ensure Docker is installed and running on your machine
2. Click Deploy button to start services

Troubleshooting

Target: Remote Deployment (R1100) {#warehouse_remote type=remote config=devices/warehouse_remote.yaml default=true}

Deploy to reComputer R1100 edge device.

Wiring

1. Connect R1100 to power and ethernet, ensure it's on the same network as your computer
2. Enter IP address reComputer-R110x.local (or check your router)
3. Enter username recomputer, password 12345678
4. Click Deploy and wait for installation to complete

Troubleshooting

Step 2: Configure Warehouse System {#warehouse_config type=manual required=true}

After deployment, open the warehouse system to complete initial setup:

1. Open browser and visit http://server-ip:2125 (use localhost for local deployment)
2. First visit will show "Set Administrator" dialog, fill in details and confirm
3. Click "Inventory List" on the left to import existing inventory (Download Excel Template)

Troubleshooting

Step 3: Jetson Local AI {#jetson_ai type=docker_deploy required=true config=devices/llm_jetson.yaml}

Deploy local LLM and TTS services on the Jetson device.

Target: Remote Deployment (Jetson) {#jetson_remote type=remote config=devices/llm_jetson.yaml default=true}

Wiring

1. Connect Jetson (e.g., reComputer J4012) to power and ethernet
2. Enter Jetson IP address and SSH credentials
3. Select model (Qwen3-8B recommended, requires ~4.3GB VRAM)
4. Choose deployment method (offline package recommended)
5. Click Deploy and wait for image import and service startup

Troubleshooting

Step 4: Voice AI Service {#voice_service type=docker_deploy required=true config=devices/xiaozhi_remote.yaml}

Deploy voice AI service on R1100, connecting to Jetson for inference. Select "Edge Computing" mode and enter the Jetson IP address (auto-filled if Jetson was deployed in previous step).

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

Wiring

1. Ensure Docker is installed and running
2. Click Deploy button to start services

Target: Remote Deployment (R1100) {#voice_remote type=remote config=devices/xiaozhi_remote.yaml default=true}

Wiring

1. Enter R1100 IP address and SSH credentials
2. Click Deploy and wait for installation to complete

Troubleshooting

Step 5: Configure Watcher Device {#watcher_config type=manual required=true}

Connect your Watcher to the local voice service:

1. Power on Watcher, connect your phone to the WiFi hotspot displayed on the device screen
2. In the popup page, tap the "Advanced" tab
3. Change OTA URL to: http://server-ip:18003 (server IP is the device running voice service)
4. Configure WiFi (2.4GHz only), device will restart and show a 6-digit verification code
5. Visit voice service console at http://server-ip:18003 and bind the device code

Troubleshooting

Step 6: Connect to Agent {#agent_config type=manual required=true}

Add an agent in the warehouse system to let Watcher control inventory:

1. Open warehouse system, go to "Agent Configuration" on the left sidebar
2. Click "Add Agent", fill in name, enter ws://server-ip:18004/mcp in Endpoint field
3. Click "Save and Start"
4. Click "MCP Endpoint" on the agent card, refresh status - "Connected" means success

Troubleshooting

Step 7: Demo & Testing {#demo type=manual required=false}

Try these voice commands:

Check the warehouse web interface to see inventory changes after speaking.

Troubleshooting

Deployment Complete

Your fully offline warehouse system is ready!

Access points:

• Warehouse System: http://<server-ip>:2125
• Voice Service Console: http://<server-ip>:18003
• LLM Health Check: http://<jetson-ip>:8000/health

100% offline operation - no internet required after deployment.