A speech-to-text app turns recorded or live speech into written text using a speech recognition model.
In plain words
You feed in audio, such as a voice note or a lecture recording, and a model writes out what was said. Open models such as OpenAI's Whisper were trained on a large, varied set of audio and can handle many languages, translate speech into English and detect which language is being spoken. A simple app wraps the model with a way to upload or record audio and shows or saves the transcript.
A quick example
Hugging Face's task page shows a short clip being turned into a line of text about travelling slushy rural roads to speak to audiences. The same idea powers automatic captions for videos and livestreams.
Where you'll meet it
- Captions for videos and live streams.
- Voice notes, dictation and voice assistants.
Related
Sources
- DocsAutomatic Speech Recognition, Hugging Face Tasks · read 29 Sept 2026
- Docswhisper, OpenAI (GitHub) · read 29 Sept 2026
Checked by WhatIsAI against the sources above · 29 Sept 2026. Spotted an error? Report it.