Thanks — great question. LiveKit Inference already supports ElevenLabs TTS standard models, but custom/community voices (including voice cloning and new conversational voice models) are not yet available through LiveKit Inference. If ElevenLabs’ new conversational modal is provided as a custom or community voice, it will not automatically appear in LiveKit. You have two options right now: • If the new model is a standard ElevenLabs model that LiveKit has integrated, it will become available once the integration is added — you’ll need to wait for LiveKit to add support. • If you want immediate access, install and use the ElevenLabs plugin and authenticate with your own ElevenLabs API key; that lets you use custom/community voices (including voice-cloned or newly released conversational models) independently of LiveKit Inference. If you’d like, tell me which ElevenLabs model name you’re looking at and whether it’s listed as a “custom/community” voice — I can confirm which path will work best for you.