The system is not returning the call analysis (success evaluation or call summary), causing all calls to be perceived as a failed case. This is not the first time we are posting about this. Would appreciate any help. Thanks call id: 0199fde5-fd63-7bb5-968e-d371bd281879

Hi there, Thank you for your message. Our team is currently out of the office. We operate Monday through Friday, from 9:00 AM to 8:00 PM Pacific Standard Time (PST). We’ll get back to you as soon as possible during our normal business hours. If your message is urgent, please mark it accordingly or include “URGENT” in the subject line, and we’ll do our best to respond promptly. Warm regards, Vapi Customer Support Team

Suggestion: Check that your assistant is configured with an

analysisPlan

and that your webhook endpoint is set up to receive the

end-of-call-report

message. In your assistant configuration, ensure you have:

js
server: {
  url: "https://your-server.com",
  timeoutSeconds: 20
},
serverMessages: ["end-of-call-report", "status-update"],
analysisPlan: { enabled: true, ... }

After updating, test a call and verify in the Vapi dashboard under Call Logs that the analysis summary and success evaluation are present. For more details, see the [call analysis documentation](https://docs.vapi.ai/assistants/call-analysis.mdx). Sources: - [Call Analysis](https://docs.vapi.ai/assistants/call-analysis.mdx) - [Debugging voice agents](https://docs.vapi.ai/debugging.mdx) - [Call queue management with Twilio](https://docs.vapi.ai/calls/call-queue-management.mdx)

I am facing the same issue. My summary does not appear. It used to be there, but today it's empty and I haven't changed anything.

@Marcello @Ian Follow this instruction to have your issues fixed..

Where to enable analysisPlan?

@Marcello Alright for the server messages try enabling “Transcript(final)” also and disable every single thing on the client messages and test to see if that works, after activating these features don’t reload the tab.. Test it again first lets see..

Okay so after turning off all client messages, and turning on Transcript (final) and end of call report (which was already there) it works! Thanks @GIDEON

@Marcello Can you paste here how you enabled transcript(final)? or @GIDEON How would i do that in the server sdk i am getting this suggestion from the types 'transcript[transcriptType="final"]' is that correct?

Hi Ian, To automatically summarize and evaluate calls, you may want to ensure the call analysis feature is properly set up. Here's how you can customize and verify it: 1. **Default Settings**: Call analysis automatically starts after a call ends, using the default prompts and schemas. 2. **Customization**: You can customize: - **Summary Prompt**: Override with your own text using

{"summaryPrompt": "Your custom text"}

. - **Structured Data Prompt/Schema**: Customize if specific data extraction is needed. - **Success Evaluation Prompt/Rubric**: Define call success criteria using scales like

NumericScale

or

Checklist

. For details, see [Call Analysis Documentation](https://docs.vapi.ai/assistants/call-analysis). 3. **Check Documentation**: Make sure your assistant’s

analysisPlan

is correctly configured. Incorrect configuration could prevent call analysis from being generated. If calls are completing without analysis data, checking your API requests for proper setup or any errors might provide further insights.

@Steyt Check the messaging sevtion on you vapi dashboard, click then scroll right to server messages, disable all variables aside end-of-call report then also enable transcript(final) from the search tag. https://cdn.discordapp.com/attachments/1429550733111459900/1430276156703379486/vapi6.png?ex=68f92ff9&is=68f7de79&hm=85b8b5b80c876ba26c96d28594932f29f3987f4996c8906805b67b6ff382be20&

I know it can be done via the dashboard but i meant from the server sdk in code

Hi Steyt, To update your assistant configuration directly through the Server SDK, you can follow this example:

import { VapiClient } from '@vapi-ai/server-sdk';

const client = new VapiClient({
  token: process.env.VAPI_API_KEY!,
});

async function updateToMinimalMessages(assistantId: string) {
  try {
    console.log('Updating assistant to minimal server messages...');
    
    const assistant = await client.assistants.update(assistantId, {
      serverMessages: [
        'end-of-call-report',
        'transcript[transcriptType="final"]',
      ],
    });
    
    console.log('✓ Success! Server messages are now:');
    console.log('  - end-of-call-report (sent when call ends)');
    console.log('  - transcript[transcriptType="final"] (only final transcripts)');
    console.log('\nAll other server messages are disabled.');
    
    return assistant;
  } catch (error: any) {
    console.error('✗ Error updating assistant:', error.message);
    throw error;
  }
}

// Usage
const assistantId = 'your-assistant-id-here';
updateToMinimalMessages(assistantId);

Explanation This script: - Uses the Vapi Server SDK to update an existing assistant. - Restricts

serverMessages

to only: -

end-of-call-report

– triggered when a call completes. -

transcript[transcriptType="final"]

– sends only the finalized transcript, not partials. - Disables all other server messages to ensure minimal and stable webhook traffic. You can replace

"your-assistant-id-here"

with your assistant’s actual ID and run it once to apply the configuration.

Thank you!

Just add the clientMessages array on the end of the const

const assistant = await client.assistants.update(assistantId, {
      serverMessages: [
        'end-of-call-report',
        'transcript[transcriptType="final"]',
      ],  // dont forget the comma
      clientMessages: [], // add this thing too
    });

This will clear out the default client messages. Separately, our team is investigating the issue where assistants only work properly after a manual re-publish. We’ll update you as soon as we have more information on that.

Perfect thanks, looking forward to a solution

We’ll share an update as soon as more information becomes available. We appreciate your understanding and patience as we continue our investigation.