Hello, tech enthusiasts! Buckle up as we explore ElevenLabs’ groundbreaking step into the realm of speech-to-text technology with their latest model, Scribe.

Having recently raised a whopping $180 million, ElevenLabs is primarily recognized for its audio generation skills. Now, they’re making waves by introducing a standalone speech detection model.

Valued at $3.3 billion, this new direction allows ElevenLabs to provide competitive speech-to-text services, boasting a vast library of voices that can aid other companies.

But what’s exciting about Scribe? This innovative model supports over 99 languages right off the bat! With more than 25 languages categorized with excellent accuracy—less than 5% word error rate—Scribe is likely to become a reliable tool for businesses worldwide.

Notably, it’s claimed that Scribe even outperforms giants like Google Gemini 2.0 Flash in crucial benchmark tests. That’s a huge win!

Previously only incorporated into its AI conversational platforms, Scribe marks the first time ElevenLabs is releasing this technology independently. According to CEO Mati Staniszewski, they’re aiming to enhance understanding during conversations, not just content generation.

Moreover, Scribe isn’t just about transcription; it incorporates smart diarization features that can recognize different speakers and accurately timestamp words for transcripts and subtitles.

Currently, Scribe works with pre-recorded audio, but ElevenLabs promises a low-latency real-time version soon. For now, it’s priced competitively at $0.40 per hour of transcribed audio.

Will ElevenLabs challenge major players in the speech-to-text landscape? Only time will tell, but with such robust features, they are certainly setting themselves up for success.