Whisper is OpenAI's speech-recognition project for turning audio into text. It provides models and code for transcription, helping developers investigate a speech-to-text capability for applications and content workflows.
Make recordings searchable
Transcribe audio when meetings, interviews, or product-related recordings need text that can be read, searched, and summarized.
Build a speech-based feature
Explore the project when an application should accept spoken input or generate captions from audio, giving you a concrete reference for the recognition step.