ElevenLabs is an AI audio generation, one of the leading companies and tools in the space, launching Scribe, its own text-to-speech model.
The company recently $180M of funding competing with speech models like OpenAI’s Whisper, Deepgram, and Gladia like models.
ElevenLabs Scribe model supports over 99 languages with features like word-level timestamps, audio-event tagging, multi-speaker detection, etc.
You can use this for your meeting summaries, movie subtitles, real-time sports titles, and others. The model makes the lowest errors in Italian (98.7%) and in English it’s at 96.7%.
Creators and businesses can use Scribe directly with the ElevenLabs dashboard to upload video or audio to generate formatted transcripts.
It is a useful model they have released. You can check it to see how it works.
You can check out the announcement video for more information.
