Build your own integration to https://octoprint.org/download/ with our example classifier integration

Stopping the Spaghetti — OctoPrint + Edge Impulse

Early detection of 3D print failures using Visual Anomaly Detection and the Edge Impulse Linux SDK

This project runs entirely on the same Raspberry Pi that hosts your OctoPrint instance.
It connects your printer’s camera feed to a locally deployed Edge Impulse model that detects abnormal print behavior (like “spaghetti”) and automatically pauses or cancels the job.

👉 View full code on GitHub edgeimpulse/example-octoprint-spaghetti-detector

Good Print

Bad Print

Quick start

• Sign up for Edge Impulse - https://studio.edgeimpulse.com/signup

• Install and configure a pi to control your printer with Octoprint - https://octoprint.org/

Copy to your Pi:

• Studio > Deploy > Linux (aarch64)
• Download > model.eim
• scp model.eim pi@octopi.local:~/example-octoprint-spaghetti-detector/
• Enable API Key control (create one called OCTO_KEY)
• 

🚀 Features

• Runs a Visual Anomaly Detection (.eim) model exported from Edge Impulse Studio
• Uses the Edge Impulse Linux Python SDK for real-time inference
• Streams directly from OctoPrint’s MJPEG feed
• Pauses or cancels print jobs when anomalies persist for several frames
• 100% on-device inference — no cloud latency or connection required

Based on

@misc{early-detection-of-3d-printing-issues, author = {Kenneth}, title = {Early detection of 3D printing issues}, year = {2023}, howpublished = {\url{ https://kaggle.com/competitions/early-detection-of-3d-printing-issues}}, note = {Kaggle} }