Developer Builds AI Language Tutor That Scores Pronunciation in Real Time
A developer has created an open-source Python Flask app that functions as an interactive language tutor, using three Telnyx AI services — text-to-speech, speech-to-text, and inference — in a single feedback loop. The app plays a flashcard phrase aloud, records the user's spoken attempt, transcribes it via Whisper, and then uses the Kimi-K2.6 reasoning model to score the response as correct, close, or wrong along with a short tip. It ships with pre-built flashcard decks for Spanish and French and uses native Ultra voices per language rather than a single English voice. Unlike similar tools that only translate speech in one direction, this app is fully interactive, with the AI both speaking to the user and evaluating their reply. The project is available on GitHub and requires only a single Telnyx API key, with no external services or phone number needed.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.
Discussion (0)
Log in to join the discussion and vote.
Log in