ElevenLabs Scribe v2 launches on Replicate
TL;DR
ElevenLabs deploys Scribe v2 to Replicate. The model transcribes speech in 90+ languages with word-level timestamps, speaker diarization for up to 32 speakers, audio event tagging, and keyterm biasing.
What dropped
ElevenLabs released scribe-v2 on Replicate. Handles transcription in 90+ languages with diarization for up to 32 speakers.
What it can do
- •Transcribes speech in 90+ languages with high accuracy
- •Delivers word-level timestamps for precise alignment
- •Performs speaker diarization identifying up to 32 speakers
- •Tags audio events like laughter or applause
- •Applies keyterm biasing to boost specific terms
What it replaces
Alternative to OpenAI Whisper for multilingual transcription needing strong diarization and event detection.
Who this matters for
- Vibe Builders: Integrate high-fidelity speaker diarization into your creative projects via the Replicate API.
- Developers: Replace standard Whisper pipelines with Scribe v2 to gain word-level timestamps and event tagging.
Harsh’s take
ElevenLabs is aggressively moving beyond voice synthesis to capture the transcription market. By launching Scribe v2 on Replicate, they bypass the friction of proprietary cloud portals and force a direct comparison with OpenAI Whisper. The inclusion of 32-speaker diarization and specific audio event tagging makes this a functional upgrade for complex meeting analysis or media production workflows.
Developers who rely on basic transcription will find the word-level timestamps and keyterm biasing particularly useful for building tighter search and indexing features. This release signals that ElevenLabs intends to own the entire audio processing stack rather than just the output side. If the accuracy holds up against established open source models, the barrier to entry for building high-end audio intelligence tools just dropped significantly.
Expect rapid adoption from teams tired of managing their own Whisper infrastructure.
by Harsh Desai
More AI news
- FeatureGoogle Launches Preferred Sources Feature for Search Results
Google introduced Preferred Sources, a feature letting users manually select preferred news outlets in search. It aims to prioritize quality journalism amid AI-generated content.
- LaunchClawTick launches on Product Hunt with one-command cron jobs for AI agents
ClawTick launches on Product Hunt. It enables cron jobs for AI agents with one command and zero infrastructure.
- FeatureWeek 2 Musk-OpenAI trial: OpenAI responds, Zilis says Musk tried to poach Altman
OpenAI responded in week 2 of its trial with Elon Musk as his suit motivations faced scrutiny. Shivon Zilis testified Musk attempted to poach Sam Altman.