ElevenLabs News 10/21/2026 AI Rating: Medium

ElevenLabs adds TTS audio effects, voice collections API and ticket merging

#ElevenLabs#TTS#API#Agents

ElevenLabs’ changelog for October 18-21, 2026 lists updates across Text to Speech, Speech to Text, Voices and its ElevenAgents platform, led by a new audio effects option for generated speech.

Details

  • Audio effects: Text to Speech and Text to Dialogue accept an audio_effects object covering filter presets, distance, reverb, background noise and panning. A new catalog endpoint lists the valid IDs. Effects do not change request cost.
  • Realtime Speech to Text: the WebSocket adds a turn_prediction commit strategy, which commits a segment when the model predicts the speaker has finished.
  • Voice collections: collections can be created, listed, updated and deleted via the API, and voices can be added or removed (six new endpoints).
  • ElevenAgents: up to 50 open tickets from the same agent can be merged into one; workflow end nodes accept an outcome of success or failure; new agents have call queueing on by default; tool_ids_strict makes an overridden tool list the complete set.
  • SDKs: JavaScript and Python SDKs v2.72.0 and CLI v1.5.0 add the new options.

What happened next

Built-in effects mean developers can shape the sound of generated speech without a separate audio-processing step, and turn prediction targets lower-latency voice agents. The changelog does not give further detail on how the effects sound or how accurate turn prediction is.