Thanks — I’m responding to your question about increasing the tool timeout when using Ultravox realtime models. Ultravox’s realtime plugin does not expose a configurable tool timeout. You cannot change a model-level "tool timeout" there. What you can do: • Handle timeouts inside your tool implementation (recommended). For example, wrap external HTTP calls with a timeout (e.g.,

asyncio.wait_for()

in Python or a fetch/axios timeout in Node.js). • Return a concise fallback or

time_exceeded_message

from your tool when the timeout triggers so the agent can continue the session. • Keep external calls non-blocking and short-lived to avoid locking the realtime connection. If you’d like, tell me what language/runtime your tools use (Python, Node.js, etc.) and I’ll provide a short code example showing how to add a timeout. Sources: Ultravox plugin guide | LiveKit Documentation | Amazon Nova Sonic integration guide | LiveKit Documentation | Azure OpenAI Realtime API and LiveKit | LiveKit Documentation | Function tool definition | LiveKit Documentation, with interactive elements