SDK releases: JavaScript SDK v2.72.0 - Added clients for voice collections
By Steven Van ·
Python gains iterable audio playback, while Scribe adds document context and conversation widgets gain carousels and end confirmation.
The JavaScript and Python SDKs for ElevenLabs both reach v2.72.0, adding clients for voice collections, the audio effects catalog and triage ticket merging. The October 12 release notes also cover updates to the CLI, Scribe packages and conversation widgets.
Voice collections, audio effects and ticket search
Both SDKs support managing voice collections, including creating and updating collections and adding or removing voices. Text to Speech and Text to Dialogue methods now accept an optional audio_effects object for filter presets, perceived distance, reverb, looped background noise and stereo panning. The catalog client provides the valid preset, environment and background noise IDs.
Triage ticket list methods gain search and filters. Search matches titles, descriptions, comments and turn comments without regard to case, with a search_match excerpt showing where a result matched. Ticket merging combines up to 50 open tickets from the same agent into one target ticket.
Realtime transcription and Python audio handling
The realtime Speech to Text clients add the turn_prediction commit strategy. Alongside silence-based segmentation, it commits a segment when the model predicts that the speaker has finished their turn.
In Python, play() and save() now accept any iterable of audio chunks, including lists and tuples. RealtimeEvents.CLOSE handlers receive a dictionary containing the WebSocket close code and reason. Handlers without parameters continue to work.
Scribe context and package updates
@elevenlabs/client@1.27.0 and @elevenlabs/react@1.17.0 add previousText to Scribe.connect() and useScribe() in microphone mode. Existing document text can be sent with the first audio chunk to provide context for casing, punctuation and sentence continuation. Output resampling also uses the same custom libsamplerate bundle path as audio input.
- ElevenLabs CLI v1.5.0 adds regenerated commands for voice collections, the audio effects catalog, triage ticket merging and agent merge proposals.
- @elevenlabs/react-native@1.2.29 updates its client and React package dependencies to these releases.
- The widget core and embed packages, both at v0.19.0, add an optional end confirmation dialog with customizable text and carousel cards containing images, text and actions. Carousels remain in the transcript without their actions after a reply. Rich content also appears during voice calls when the transcript is visible.
