Open project
RF Connector AI
Identify RF connectors from your phone camera. Open dataset. On-device classifier. Mobile-first.
About the project
What is RF Connector AI?
RF Connector AI (RFCAI) is an open training-data project for fine-grained visual classification of radio-frequency connectors — the SMA, N-type, BNC, TNC, SMP, and other small hardware pieces that live on the end of RF cables. The goal is a lightweight, on-device model that runs in real time on a smartphone camera, so technicians in the field can identify an unfamiliar connector in seconds.
The dataset is collected via a mobile-first Flutter app and a web-based labeling tool. Anyone with access can contribute images, review labels, and trigger training runs. The current on-device classifier is exported to ONNX and runs via Unity Sentis on iOS and Android. A public Streamlit demo is available at /demo/. To contribute images or help review labels, sign in or get in touch with chris@aired.com.