Scribe v2

Scribe v2 is ElevenLabs' advanced transcription model designed for batch transcription, subtitling, and captioning at scale. It enhances the stability and accuracy of its predecessor, Scribe v1, by effectively handling long-form audio, pauses, tone changes, and extended silences. The model supports over 90 languages, delivering reliable transcripts across diverse speakers, accents, and delivery styles.

Scribe v2 is ElevenLabs' advanced transcription model designed for batch transcription, subtitling, and captioning at scale. It enhances the stability and accuracy of its predecessor, Scribe v1, by effectively handling long-form audio, pauses, tone changes, and extended silences. The model supports over 90 languages, delivering reliable transcripts across diverse speakers, accents, and delivery styles.

• 01Keyterm Prompting for context-aware transcription
• 02Built-in entity detection with precise timestamps
• 03Automatic multi-language transcription
• 04Smart speaker diarization for clear speaker labeling
• 05Dynamic audio tagging that detects non-speech events