Hi Vapi Team, I'm currently testing my outbound caller setup and have encountered an issue that consistently blocks the assistant from checking calendar availability during a call. Each time I initiate a test call (acting as the prospect), the assistant reaches the point where it should run the getTimes tool — but instead, there is complete silence. Upon inspecting the assistant configuration afterward, I’ve noticed that the getTimes tool has been removed. Even after manually re-adding the tool and publishing the assistant, the issue repeats: as soon as I update the VAPI Assistant using an API call from my Make.com scenario, the assistant no longer has access to the tool. As a result, any attempt to book a demo fails silently. Has anyone else experienced this behaviour? Is there a recommended way to ensure tools persist through assistant updates via the API? Any insight would be greatly appreciated — thanks in advance!

Hi Lorenzo! 👋 I understand you're facing an issue where your getTimes tool is being removed from your assistant after API updates from Make.com, preventing calendar availability checks during outbound calls. While I couldn't find specific information about the getTimes tool itself, I can definitely help with the tool persistence issue after API updates. When updating a VAPI Assistant via API calls (like from Make.com), there are several best practices to ensure tools persist: 1. Include all existing properties in your API update request - When making a PATCH request to update your assistant, you must include all existing properties, especially within nested objects like the

model

property. According to [Discord conversations](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1333975440003563604), omitting existing properties can cause them to revert to default values or be removed entirely. 2. Use the correct structure in your request body - Make sure your API request includes the

toolIds

array within the

model

section. Here's an example of a properly structured PATCH request:

json
{
  "model": {
    "provider": "openai",
    "model": "gpt-4o",
    "toolIds": [
      "your-tool-id-here"
    ]
  }
}

3. Use tool IDs instead of tool names - According to [user reports](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1338953413886480425), ensure you're using the correct tool ID to enable tools, as using the tool name might result in a new, transient tool being created instead of correctly referencing your existing tool. 4. Check your Make.com scenario configuration - Double-check the configuration of your Make.com module to ensure you've entered the correct information without typos or extra spaces, as mentioned in [troubleshooting discussions](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1334889631333875793).

It's important to ensure that the tools array is not omitted, as doing so could lead to the removal of previously configured tools, such as gettimes.

Thanks a lot for your help - the problem seems to be solved now and the tool is running correctly.