Hello,

I’ve been testing a modified version of main.py located in depthai-experiments/neural-networks/object-detection/spatial-detections/. I successfully updated the code to run with my custom .tar.xz model, and the app builds and runs properly using:
oakctl app run .

The build completes without errors, and the app deploys as expected.

However, when I open the DepthAI Visualizer in the browser (replacing localhost with the camera’s IP on port 8082), I’m stuck on the screen that says:

To isolate the issue, I also tested an unmodified example (social-distancing) and followed the same steps. Although the build was successful, I encountered the exact same issue when trying to view the visualizer — stuck on the same WebSocket connection message.

Given that both a modified and an unmodified example exhibit this problem, I'm wondering:

Is this issue likely to be on my end (e.g., browser, firewall, or networking)?

Or could there be a problem with the WebSocket server or DepthAI Visualizer on the device?

Any guidance would be appreciated!

Best regards,
Badri Pratti

    BadriPratti
    We'll get back to you with a solution. Thanks for reporting.

      BadriPratti
      A fix has been pushed to depthai v3_develop branch.
      What you need to do is put

      --extra-index-url https://artifacts.luxonis.com/artifactory/luxonis-python-snapshot-local
depthai==3.0.0a15.dev0+adc7ca692eef7dae97593176cb1f977eed6f8837

      in the requirements.txt file. Then re-run the app.

      Thanks,
      Jaka