Real-time printer dashboard with vision monitoring
A self-hosted dashboard that streams a printer camera over WebRTC, telemetry over MQTT, and runs a local vision model watching for failures, with no cloud dependency.
An industrial-grade 3D printer ships with a cloud app and a camera you can only watch through the vendor. We wanted a local screen that showed the camera, the telemetry, and a judgment on whether the print was failing, without sending anything off the network.
What we built
A WebRTC camera stream relayed locally, so the feed loads instantly on any device on the network
Telemetry over MQTT, pushed to the browser through one WebSocket
A local vision model that inspects frames and raises a flag on likely failures
Graceful states for a missing camera or a missing model, so the dashboard never lies
A phone layout for checking on a print from another room
What carries over
The pattern generalizes to any machine with a camera and a data feed: keep the data local, show one screen, and let a model watch so a person does not have to.
Result
One screen, on the local network, that shows what the machine is doing and flags a failing print.
codeaspen.com/work/printer-vision-dashboard
You dream, we deliver.
Bring us the problem. We’ll scope the fix.
A 20-minute scoping call: what you have, what is breaking, and what a fixed-scope build would look like. If it is not a fit, you will know by the end of the call.
