Hi there! I'm running into a few issues that I'm hoping to get help with 1) “It seems like Vapi allows us to name the MCP server itself, but not the tools within the MCP server. That makes sense. However, my agent is continually trying to call the MCP server itself (CL_mcp_tool) which is not callable and so causes an error. (see screenshot 1) 2)Additionally it seems like we can only configure start/end message for the MCP server itself, not the tools within it, which doesn’t make sense to me. Am I missing a way to configure start/end messages for the tools WITHIN the MCP server? 3) When my agent calls the send_text tool first, it then fails to call my tool from my MCP server and just goes silent when it attempts...Is there a way to fix this? https://cdn.discordapp.com/attachments/1379172027457605732/1379172027696676985/Screenshot_2025-06-02_at_11.56.41_AM.png?ex=683f458b&is=683df40b&hm=efaa42342b001dbba400d84f48ce61a807fc63a1f0d22b2d306a793224484906&

Hi Xtina277, For your issues with the MCP server and tool calls: 1. **Tool Call Errors**: If the agent is not reliably calling the correct tools, ensure that your endpoint is returning JSON in the format expected by Vapi. It would be useful to log the incoming requests from Vapi to see if there's a mismatch in expectations. Additionally, verify that tool names are unique within your MCP server configuration, as this can cause errors when tools are misidentified. 2. **Configuring Request Start Messages**: To configure custom messages such as "request_start" for tools within your MCP server, you can set specific message configurations in the MCP tool setup process. This involves defining what you want the agent to say before initiating the tool call. However, if the default message is not editable, consider using an external Custom Tools configuration via the [Vapi Dashboard](https://docs.vapi.ai/tools/custom-tools) for further customization. 3. **Send Text Tool Failure**: Ensure that the tool is correctly configured and that all required parameters are being passed. Double-check the tool's setup in your MCP server and the Vapi dashboard for accuracy. If there's an API response indicating failure, this often means that either the tool isn't setup properly or a required parameter is missing. If issues persist, review logs for more detailed error information and ensure that all your Vapi configurations align with [Vapi MCP Server](https://docs.vapi.ai/sdk/mcp-server) requirements.