Thanks — great question. You're asking whether you can update STT or switch STT providers during an active live agent session. Short answer: yes. What you can do • Update STT behavior on the fly by calling the STT instance's

update_options()

during the session. For example, you can change language or other provider-specific options while the call is ongoing to adapt recognition in real time. • Switch between different STT providers by performing an agent handoff (transfer the conversation to an agent configured with a different provider) or by reconfiguring the session to use the other provider mid-session. Notes and things to consider • When using

update_options()

, you can change settings such as language to improve recognition without reconnecting the session. • Switching providers may require reinitializing provider-specific connections or credentials; using an agent handoff is a common pattern to move the conversation to a differently configured environment. • Plan for brief interruptions or latency during provider transitions (reconfiguration or handoff) since audio routing/processing may change. If you’d like, tell me which STT providers and which options you want to change (for example: language, model, or provider swap), and I’ll give you a short example of the code or steps to use. Sources: ElevenLabs Change Language | LiveKit Documentation | Agents and handoffs | LiveKit Documentation