I would guess that they're using OpenAI's Whisper, which is open source: https://github.com/openai/whisper https://github.com/openai/whisper It does speech-to-

I would guess that they're using OpenAI's Whisper, which is open source: https://github.com/openai/whisper https://github.com/openai/whisper It does speech-to-text, then you can use the full force of all the text analysis tools that are out there.