Introducing Scribe v2

Scribe v2 is the latest transcription model from ElevenLabs, offering enhanced accuracy and stability for transcribing, subtitling, and captioning across more than 90 languages. It significantly improves upon its predecessor by better handling long-form audio, diverse speaker accents, and complex recordings while achieving the industry's lowest word error rate. Key features include context-aware transcription through keyterm prompting, built-in entity detection with precise timestamps, and automatic multi-language transcription without the need for manual segmentation. Additionally, Scribe v2 provides enterprise-level features such as smart speaker diarization, dynamic audio tagging, and compliance with multiple data protection standards, making it suitable for a wide range of applications in marketing, media, research, training, and compliance. Available through ElevenLabs Studio and API, Scribe v2 enables developers and enterprises to automate audio workflows and manage large volumes of content securely and efficiently.