The device turns speech into text and sends the results to your retail voice dashboard.
The device turns speech into text and sends the results to your retail voice dashboard.
Enter the server address and login details, then deploy. Keep its dashboard address and access key for the next step.
| Issue | Solution |
|---|---|
| Cannot connect to the server | Check the server address and SSH login details, and that the server is reachable from this machine |
| Deploy stops while waiting for the service to turn healthy | Sign in to the server and run docker compose logs voice-service; MySQL or MinIO may still be starting |
| Dashboard does not open afterwards | Confirm the voice-web container is running and port 3000 is reachable |
Deploy to a Linux server the recording devices can reach.
recomputerChoose your device model and connect the microphone. Enter its login details, dashboard address and access key, then deploy.
| Issue | Solution |
|---|---|
| Cannot connect to the device | Check the device IP and SSH login details |
| Microphone not detected during deploy | Move the reSpeaker microphone array to a USB-A host port and run the deploy again |
| No transcription reaches the dashboard | Check the dashboard address and access key match what Step 1 issued |
Choose the model that matches your device.
recomputerzhPrimary language of the transcript. SenseVoice on this board covers Chinese and English; pick the one spoken most in the store.
1ALSA card number of the reSpeaker XVF3800. Run `arecord -l` on the device; the array shows as ArrayUAC10.
Speak into the microphone, then open the retail voice dashboard. Confirm that the recording and its text appear.
| Issue | Solution |
|---|---|
| The recording does not appear | Check the recording device is powered and the microphone is connected, then record again |
| The recording appears without text | Check the speech service container is running on the recording device |
Clip records the conversation, your phone uploads the audio, and the retail voice platform turns it into text.
Choose your server model, enter its address and login details, then deploy. The dashboard and transcription service are installed together.
| Issue | Solution |
|---|---|
| Cannot connect to the server | Check the server address and SSH login details, and that the server is reachable from this machine |
| Deploy stops while waiting for the service to turn healthy | Sign in to the server and run docker compose logs voice-service; MySQL or MinIO may still be starting |
| The phone app cannot reach the transcription endpoint later | Confirm the capture-gateway container is running and port 18621 is reachable |
Choose the model that matches your device.
recomputerlocal_whisper.http://<deployment-device-IP>:18621; enter the access key provided by the platform administrator.| Issue | Solution |
|---|---|
| Test Connection fails | Check the Base URL is http://<deployment-device-IP>:18621, that the phone can reach that address, and that port 18621 is not blocked |
| Access key rejected | Re-enter the access key exactly as provided by the platform administrator |
| Clip does not pair | Follow the Clip pairing instructions linked above |
Open the dashboard and sign in to view and manage recordings.
| Issue | Solution |
|---|---|
| Page does not load | Check the platform deployment finished successfully and the voice-web container is running and its port reachable |
| Cannot sign in | Confirm the dashboard address and the login details from the platform deployment |
Record a short clip and upload it from your phone. Confirm that the recording and its text appear in the dashboard.
| Issue | Solution |
|---|---|
| The upload does not appear in the dashboard | Confirm Test Connection succeeded in Step 2 and the app configuration was saved |
| The recording appears without text | Check the speech service container is running on the server |