Add Techlomedia as a preferred source on Google.
ElevenLabs has announced the launch of Scribe v2 Realtime, its most advanced Speech-to-Text model yet. The new model can deliver live transcription in under 150 milliseconds while maintaining top-tier accuracy across more than 90 languages, including 11 Indian languages such as Hindi, Tamil, Telugu, Malayalam, Bengali, Gujarati, Kannada, Odia, Marathi, Punjabi, and Sindhi.
The company says Scribe v2 Realtime sets a new benchmark for real-time multilingual communication. It allows developers and enterprises to build faster, more natural, and inclusive voice experiences across sectors like customer engagement, healthcare, media, education, and live streaming.
According to ElevenLabs, the model achieves 93.5% accuracy on the FLEURS benchmark, which covers 30 European and Asian languages.
Scribe v2 Realtime is designed for developers and enterprises creating voice assistants, live meeting tools, and real-time captioning systems. It brings human-level understanding to live environments, enabling instant responses with impressive precision.
The model supports several advanced features such as negative latency prediction, text conditioning, voice activity detection (VAD), and manual commit controls for fine-tuned performance during live streaming or transcription.
Enterprise use cases include real-time transcription of customer calls, live medical dictation, meeting transcriptions, media captions, and accessibility tools for education. ElevenLabs has also emphasized its commitment to data localization, offering India-based data residency options to help organizations comply with local data protection regulations.
Scribe v2 Realtime can also integrate with ElevenLabs Agents, which allows developers to build conversational AI systems that sound more natural and human-like. These systems can be used for support, sales, and in-product experiences.
Scribe v2 Realtime is now available through the ElevenLabs API, and developers can access it directly via ElevenLabs Agents. This update brings instant, high-quality transcription capabilities to real-world applications in multiple industries.
For more information, visit elevenlabs.io/docs/capabilities/speech-to-text.
Affiliate Disclosure:
This article may contain affiliate links. We may earn a commission on purchases made through these links at no extra cost to you.
