ElevenLabs releases Scribe v3 series transcription models and launches a medical version
Understand this story
ElevenLabs has released two speech-to-text models, Scribe v3 and Scribe v3 Live, aimed at recordings and real-time audio respectively. ElevenLabs says Scribe v3 has 27% fewer word errors than Scribe v2 across 102 languages, with the highest speaker-labeling accuracy among the models it has tested; Scribe v3 Live has 43% fewer errors than Scribe v2 Realtime. Scribe v3 Live is already available in ElevenAgents, and both models are now in ElevenAPI, with switching from Scribe v2 requiring only a one-line model ID change. Both models run in the US by default; for enterprise customers with data residency needs, isolated EU, India, and Singapore environments are also offered. ElevenLabs subsequently launched Scribe v3 Live Medical in ElevenAgents, bringing medical transcription to real-time calls. This version is aimed at enterprise customers who have signed a Business Associate Agreement (BAA) and enabled zero-retention mode, and is HIPAA-eligible; patient data is stored in the US by default, and enterprises needing overseas storage can use the isolated EU, India, or Singapore environments.
AI overview based on source reports · Updated 6 hr ago
Developments
2 developments
- Nov 17 20:00 · 1 reportsElevenLabs releases Scribe v3 Live Medical real-time medical transcriptionElevenLabs Blog:Scribe v3 Live Medical helps conversational agents get clinical details right
- Nov 10 20:00 · 1 reportsElevenLabs releases Scribe v3 and v3 Live speech-to-text modelsElevenLabs Blog:Introducing Scribe v3 and v3 Live, our most accurate speech to text models
Coverage timeline
Follow the coverage to explore different perspectives.
- ElevenLabs BlogScribe v3 Live Medical helps conversational agents get clinical details right
ElevenLabs released Scribe v3 Live Medical, a real-time Speech to Text model built on Scribe v3 Live and optimized for clinical language, available now in ElevenAgents. In internal tests it had the lowest WER among medical models on PriMock57 (12.8%) and the lowest semantic WER on Fareez OSCE (0.6%), while matching Scribe v3 Live on everyday speech and about 150 ms latency. It supports up to 10,000 custom terms, waits for callers to finish numbers, handles more than 90 languages, and is HIPAA-eligible for enterprise customers with a BAA and Zero Retention Mode enabled.
- The DecoderElevenLabs Scribe v3 Live uses conversational context to stop voice agents from cutting off callers mid-sentence
ElevenLabs released Scribe v3 for pre-recorded audio and Scribe v3 Live for low-latency real-time use, both available through the ElevenAPI and ElevenAgents platform. The company says Scribe v3 cuts word errors by about 25 percent versus its predecessor across 102 languages, while Artificial Analysis benchmarks confirm top-tier accuracy but note slower speed and mid-pack pricing. Scribe v3 Live analyzes conversational context alongside the audio signal to detect when a speaker has actually finished, and both models support preloaded specialized terms, natural-language formatting rules, and mid-sentence language switching.
- ElevenLabs BlogFeaturedIntroducing Scribe v3 and v3 Live, our most accurate speech to text models
ElevenLabs launched Scribe v3 for recorded audio, which makes 27% fewer word errors than Scribe v2 across 102 languages and has the lowest diarization error of any model tested, and Scribe v3 Live for realtime audio, which makes 43% fewer semantic word errors than Scribe v2 Realtime and adds speaker diarization and improved turn detection. Scribe v3 Live is available today in ElevenAgents, and both models are available in ElevenAPI, where switching from Scribe v2 is a one-line change to the model ID (scribe_v3 or scribe_v3_live). Both run in the US by default with isolated EU, India, and Singapore environments, Zero Retention Mode, and SOC 2, ISO 27001, PCI DSS Level 1, and HIPAA-eligible coverage.
Activity for this story
- Current comparable activity
- 7
- Peak comparable activity
- 8Nov 18 03:00
- Change in the past 24 hours
- –
Trends compare the same fully observed participants, which may cover fewer participants than current activity. Move the pointer or tap the chart to inspect hourly activity; use the left and right arrow keys with a keyboard.