ElevenLabs Scribe

ElevenLabs Speech to text is a speech-to-text model that specializes in converting speech into text with remarkable accuracy across multiple contexts and languages. It houses two main features namely, Scribe v2 and Scribe v2 Realtime. The former focuses on the transcription of audio and video content into text, perfect for creating captions, subtitles, and editable transcripts for various forms of recorded content. It stands out for its ability to accurately transcribe specific words based on context, marked sound events in transcripts, and distinguish and label every speaker in a dialogue. The latter, Scribe v2 Realtime, is designed for real-time applications with an emphasis on things like live calls, meetings, or AI agents requiring immediate transcription. It uses a streaming-first architecture to provide real-time results while still maintaining accuracy. It also includes features like precision speech segmentation for smoother live processing and voice activity detection. Both versions of Scribe support over 90 languages and can be incorporated into your products using their API.