Agent not continuing after using tool
# support
u
Hello VAPI Support Team, I'm encountering an issue with my VAPI AI agent that's integrated with an N8N webhook. The webhook returns data in this format: call ID:
f21195ab-9bef-4e81-8b12-4844818f7b01
at sec 00:57 it got the data but never talks until I asked again about the AvailableTime.
Copy code
json
{
   "results": [
     {
       "toolCallId":  "call_LbjUaWAd14gHJksWZJEiBOj9",
       "result": "AvailableTime: 9-11, 2-4"
     }
   ]
}
I’ve confirmed via the VAPI console that the data is being received correctly. However, after the data comes in, the agent doesn't proceed with the conversation automatically. When I query it about the data, it confirms that the information is present, but it doesn't use it to continue the dialogue immediately. I have verified that there are no timeout issues and the webhook response is correctly formatted. I’ve also reviewed the configuration for synchronous tool calls, tool response handling, and prompt settings. Despite all these checks, the agent still stops after receiving the data. Could you please help me troubleshoot this? Any insights or suggestions would be greatly appreciated. Thanks in advance! https://cdn.discordapp.com/attachments/1345062021741412534/1345062188636831796/image.png?ex=67c32e47&is=67c1dcc7&hm=5a792121269af2f4352fc19898e93bdbcee47b99ce3fdc1780e127b7939a3ce0& https://cdn.discordapp.com/attachments/1345062021741412534/1345062189110923315/image.png?ex=67c32e47&is=67c1dcc7&hm=226a70c16d9ddec37c775d8f02cfa1454fbe473607775a113c2cd6b6a4282e07&
v
Hi there! 😊 It sounds like you're dealing with a tricky situation with your VAPI AI agent. Here are some troubleshooting steps you can take to help resolve the issue where the agent doesn't continue the conversation after receiving data from the webhook: 1. **Verify Configuration and Integration**: - Ensure that the VAPI is correctly configured to send and receive the correct data from N8N. Double-check that the data fields match on both platforms. 2. **Check Logs**: - Access the logs on both VAPI and N8N to identify any errors or warnings that might indicate where the process is failing. 3. **Webhook Settings**: - Confirm that the webhook settings in VAPI are configured to trigger scenarios properly with the correct conditions and data setup. 4. **Dynamic Content Configuration**: - Make sure that any placeholders in your voice pipeline are correctly configured to handle and replace dynamic content. Verify that the dynamic variables are set correctly during each call. 5. **Response Format**: - Ensure that the webhook response includes correctly formatted JSON and that the
toolCallId
matches the expected format. The response should also have the correct content-type set to
application/json
. 6. **Test with Static Data**: - Try using static data in tests to isolate issues related to dynamic data handling. This can help determine if the problem lies with the data being sent. 7. **Review API and Server Configuration**: - Check for any misconfigurations in API keys, access permissions, or any recent changes that might be affecting the agent's performance. If you've gone through these steps and the issue persists, it might be worth simplifying the scenario to see if a basic setup works before gradually adding complexity back in. Let me know how it goes or if you need further assistance! 🌟 Sources: - [VAPI Documentation on Webhook Configuration](https://docs.vapi.ai/customization/webhook-configuration)
- [VAPI Troubleshooting Guide](https://docs.vapi.ai/troubleshooting)
u
I checked all... for step 5 should I add it like this:
Copy code
json
{
  "results": [
    {
      "content_type": "application/json",
      "tool_call_id": "call_sgMzU8GIPlQmtiSAhjwJJk95",
      "result": "AvailableTime: 9-11, 2-4"
    }
  ]
}
or add it in the header? anyways tried both and didnt work and here is the test call id
6c18e117-3d27-43b8-8a32-85ff4e81d2ad
and toolCallId
call_sgMzU8GIPlQmtiSAhjwJJk95
@Mason | Building KOI
@User
s
@!Mohammad Ali as we discussed over general thread, you have to remove the request-complete message type and then it will work as your expecting it to work.