Use the tool below to record yourself and see ASR feedback on your pronunciation.
The system is designed to be embedded in a mobile app for my dissertation. It uses the trellis and backtracking scores from Wav2Vec2 results outlined here.