Hey Lads, does the Real-Time-API have issues with accessing and communicating with tool calls? I was pretty stuck on my agent constantly inventing information after a tool call triggers, and it ignores the response entirely. A head scratcher until i changed LLM to 4o-mini, and it seemed to start working great again :/

yes it does have few limitations, and its from model vendor side.

@Shubham Bajaj Alright, ty Any known ways currently to use Tools with Realtime then or is it a no go?

it works but the model has few issues so its un predictable.