ElevenLabs Releases New Cue Audio Alignment Model
The company launched the model to map text characters to recorded speech across over 90 languages.
Updated on Oct. 29, 2026 in Artificial Intelligence
Live Poll
Do you trust current AI tools to provide accurate automated subtitles for your media?
ElevenLabs has introduced Cue, an audio alignment model designed to place text characters at precise points within recorded speech. The technology is now available through the company's Forced Alignment API.
Why it matters
This model improves the synchronization of text and audio, which is essential for accurate captioning and content processing. By supporting dozens of languages, it offers a scalable solution for global media production.
Cue processes one hour of audio in 0.5 seconds on a production GPU and supports more than 90 languages. It achieves an average timing error of 179 milliseconds across 29 languages, dropping to 57 milliseconds for English audio with noise.
The players
ElevenLabs
An artificial intelligence company that specializes in speech synthesis and voice technology.
The details
The model functions in a single pass to map individual text characters to specific moments in a recording. It follows the existing Scribe speech-to-text model, allowing developers to integrate the alignment capabilities directly into their own applications.
Timeline
ElevenLabs detailed the Cue model on October 28, 2026.
The Tech Race
This release follows the company's Scribe speech-to-text model, marking a shift toward more granular control in audio-text synchronization. It positions ElevenLabs to compete with advanced automated transcription and synchronization tools currently used in broadcast and digital media.
Users and developers can expect faster and more accurate automated captioning for long-form audio content. The model significantly reduces the time required to sync text, allowing for more efficient workflows in media post-production.
The takeaway
Precision in audio alignment is critical for modern content accessibility and automated transcription services. By leveraging GPUs for rapid processing, developers can significantly lower the overhead associated with large-scale media synchronization projects.
Further reading
Learn more about advancements in Artificial Intelligence to see how alignment models are evolving.
Source note: This article includes information reported by TokenPost.
Live Poll
Do you trust current AI tools to provide accurate automated subtitles for your media?
