This question originally came up in our Slack community and the thread has been consolidated here for long-term reference.
I’m using an MCP server for my tools with a LiveKit agent. After the user confirms their name, the LLM calls a tool to create an appointment. But if the user says their name multiple times and speech is being transcribed, the LLM calls the tool multiple times.
LiveKit is discarding the first tool call result and the LLM can only see the response from the second call. Has anybody faced this issue?