ElevenLabs Scribe

ElevenLabs Scribe is a speech-to-text family for transcribing recorded and live audio. Scribe v2 produces editable transcripts, captions, and subtitles for audio and video, while Scribe v2 Realtime streams transcription for agents and live applications with low latency. The models support more than 90 languages and include speaker and entity detection, word and entity timestamps, keyterm prompting, dynamic audio-event tags, sensitive-information redaction, voice activity detection, and access through ElevenLabs Studio, APIs, and SDKs.