I have a prompt containing 2 different points where I send data to Make using 2 different tools. Tool 1 is called once a data point is collected, Tool 2 is called 30-60 seconds later after more data collected. It is currently not calling any tool in some calls, or calling tools together at another call, both behaviors are not what we want. This was working fine with my last tests 3 weeks ago, now not consistent at all. Last call IDs: c89d3261-ef71-4042-8259-8cabf5def4b7: No tool calling 4ba96d7e-b1c5-4fd7-a99e-53a9f9d57d83: Calling 2 tools same time 35af7dc0-cbfb-42e2-9778-7b0d64500d8e: No tool calling d418ddd8-3072-483f-8102-dd4b21b53ee9: Older test where it was working normally

Could you please check if there have been any recent changes to your configuration or code, especially related to the timing and order of tool calls? Additionally, have you updated to the latest Vapi platform version, as there have been recent updates that might affect tool calling behavior? You might want to review the [recent updates](https://docs.vapi.ai/changelog) to ensure compatibility. Sources: - [Detailed Report on Recent Updates or Changes to the Vapi Platform Affecting Tool Calling Behavior](https://docs.vapi.ai/changelog) - [Troubleshooting Tool Execution Issues in Vapi](https://discord.com/channels/1211482211119796234/1211483291191083018/1336399188665040968)

what model are you using? And can you also share your definitions of the tools

Model : gpt-4o First tool description: The tool will send extracted variables from the call to Make scenario It should contain JSON Body with the following { "PartNumber": {{PartNumber}} } In the prompt it is triggered through: "Trigger the 'getUPS ' tool with value {{PartNumber}}" Second tool description: The tool will send extracted variables from the call to Make scenario It should contain JSON Body with the following { "PartNumber": {{PartNumber}}, "Desc": {{PartType}}, "equipment": {{Equipment}}, "email": {{email}}, "collectedPhoneNo": {{collectedPhoneNo}}, "source": {{source}}, } In the prompt it is triggered through: "call function sendTo_GHL_UPS"

yeah good enough descriptions and a good model. Seems like a lot of people are having problems but vapi for the most part has no part in the tools it's the llm calling it. May be an open AI issue, I'll look more into it while we wait for vapi support to confirm or not

Hey @Emad, can you try 1-2 calls w/o your api key?