Thiagohgl/ai-pronunciation-trainer

Uses OpenAI's Whisper for speech-to-recognition paired with Epitran-based phonemic analysis to compare spoken pronunciation against target text, scoring accuracy without external dependencies like OR-Tools. Built as a Python web application deployable locally or hosted online, supporting multiple languages through modular database and phonemic model architecture. Provides real-time feedback via a Chrome-optimized interface with editable target phrases and console-based error reporting for debugging.

452 stars. No commits in the last 6 months.