Hey, I’m having an issue with dynamic variables not loading in time during inbound calls. Here’s what’s happening: • I have a Vapi assistant with {{name}}, {{petName}}, and {{petType}} set up in the first message. • I’m using the “Server URL” (not a tool) to trigger an n8n webhook on call start. • In the webhook, I query Supabase for the customer and pet data based on customer.number, then respond with a Respond to Webhook node like this: { "assistantId": "xxxxx", "assistantOverrides": { "variableValues": { "name": "John", "petName": "Zena", "petType": "cat" } }, "customer": { "number": "+19782221234" }, "phoneNumberId": "xxxxxx" } Everything looks fine in the webhook response logs, and the data is correctly passed back. BUT… the assistant still starts the call by saying “Hi, thanks for calling” instead of using the name. However, when I manually trigger the call via API (using the same values), the assistant says the name correctly. It seems like the webhook response is arriving too late during inbound calls. Is there any way to delay the first message until the webhook returns the assistantOverrides? Or am I missing something in the config? Would appreciate any guidance!

Hi there! 👋 I understand you're having an issue with dynamic variables not loading in time during inbound calls, causing your assistant to use a generic greeting instead of the customer's name. I can't see your specific configuration beyond what you've shared, but I can definitely help with ways to delay the first message until your webhook returns the assistantOverrides. Based on the information I found, here are several approaches you can try: 1. Use

startSpeakingPlan.waitSeconds

to add a delay before the assistant starts speaking:

javascript
const assistant = {
  firstMessageMode: 'assistant-speaks-first',
  startSpeakingPlan: {
    waitSeconds: 2 // Waits 2 seconds before speaking
  }
  // ... other configurations
};

This will add a 2-second delay before the assistant starts speaking, which might give your webhook enough time to return the data ([source](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1343915991670853654)). 2. Set

assistantOverrides.maxDurationSeconds

to make the assistant remain silent for a specific period:

javascript
assistantOverrides: {
  maxDurationSeconds: 5 // Remains silent for 5 seconds
}

This ensures the assistant remains silent for the specified duration, giving your webhook time to return ([source](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1335807780287676456)). 3. Set

firstMessageMode

to

assistantWaitsForUser

to make the assistant wait for user input before responding:

javascript
const assistant = {
  firstMessageMode: 'assistantWaitsForUser',
  // ... other configurations
};

This makes the assistant wait for the user to speak first ([source](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1305065394662408192)).

Hey! To help track down this issue, could you share: - The call ID - What response you expected to get - What response you actually got instead This would really help us figure out what went wrong!

Hi, thanks for help me!

Tiago Bardini, Your dynamic variables are not being applied because your assistant overwrite is empty. That's because your phone number used doesn't have a server URL; it's using an assistant assigned to it. You can check your server for the logs that your server is not getting requests (server message of type 'assistant request'). I suggest you remove the assistant ID from your phone number, assign a server URL to your phone number, and then VAPI will trigger an assistant request server message to your server URL at this time. You can return the dynamic variables as passed off as part of assistant overwrites.

This is very confuse! can you provide documetation, something i cant take a look please. I am trying to handle inbound calls using Twilio and Vapi, but I’m facing issues with sending dynamic assistantOverrides when the call hits the Vapi webhook. My Twilio number is currently configured to point to Vapi’s webhook: https://api.vapi.ai/twilio/inbound_call I use N8N to process incoming calls. I successfully fetch customer and pet data from Supabase based on the caller’s number. This dynamic data works perfectly when I initiate outbound calls using the Vapi API, including the assistantOverrides like: • name: Tiago Bardini • petName: Zena • petType: cat However, for inbound calls, even though I’m returning a proper webhook response, the assistantOverrides are not being applied. I also tried sending XML responses and different formats, but it either loops or fails silently. I’m now considering using a transient assistant to apply these dynamic values. Could you clarify: 1. What is the recommended way to inject assistantOverrides for inbound Twilio calls? 2. Is creating a transient assistant the correct approach for this use case? I’d appreciate any insights or examples that could help me get this working properly. Thank you!

@Thiago Bardini https://cdn.discordapp.com/attachments/1353047867438137505/1355498410890760323/message.txt?ex=67e925c5&is=67e7d445&hm=03a7d9bc4b6f542852e0a718094e8e2e81691c2ab499750eae997a9541e4fbcc&

Thank you!

Marking this ticket as Solved ✅