Welcome to my central repository for all things related to 3D printing. This repo houses my 3D models, G-codes, datasets, and my custom AI-based print failure detection system built for the Creality Ender-3 V2 Neo.
Models/: My collection of 3D models and generated G-code files.src/&utils/: Source code for my real-time YOLO-based Print Failure Detection System.
One of the core projects in this repository is a real-time 3D print failure detection system using AI (YOLO). Point a webcam at your printer, and the system continuously watches for common FDM printing defects and can automatically pause or kill the print before things get worse.
Built for the Creality Ender-3 V2 Neo, but works with any FDM printer + webcam setup.
| Defect | Description | Severity |
|---|---|---|
| 🍝 Spaghetti | Print detaches from bed and extrudes into a tangled mess | 🔴 Critical — auto-stops printer |
| 🧵 Stringing | Thin filament wisps between travel moves | 🟡 Cosmetic — logs alert |
| 🔵 Zits | Small blobs/bumps on the print surface | 🟢 Minor — logs alert |
I've built an integrated web dashboard (src/print-dashboard.py) using Gradio that combines printer control and AI monitoring into a single interface:
- Smart Hardware Detection: Automatically tests COM ports to find the active 3D printer and selects the best available webcam.
- SD Card Management & Printing: Fetch G-code files from the printer's SD card and start prints remotely via USB.
- Live AI Monitoring: Streams the webcam feed with real-time YOLOv8 annotations directly to your browser.
- Native Auto-Stop: Safely shares the active serial connection to trigger an
M112Emergency Stop if critical failures (spaghetti) are detected.
How to run the dashboard:
pip install -r requirements.txt
python src/print-dashboard.pyAccess the UI at http://127.0.0.1:7860
I documented the entire process of building the failure monitoring system in a series of blog posts. Check them out for a deep dive into the development, training, and deployment: