Speech to text

Speech to text models transcribe audio input to a written output. Transcribed text is searchable, making it useful for content moderation, call center auditing, and podcast SEO. And speech-to-text models can also automatically generate video subtitles and other accessibility text. But for automated transcription to be useful, it needs to be accurate. OpenAI's open-source Whisper model delivers a generational leap in transcript quality, accurately transcribes dozens of languages, and properly handles technical terms. Try it for yourself with our Whisper demo app then deploy the model for your own projects.