This is my configurations assistant. But the assistant not read the knowledge base. This is the knowledgebaseId:

{
  "id": "62fe63f2-da01-4275-a811-4a5f89b88f27",
  "orgId": "a683a00d-fd3e-41d0-b569-3b73957bb165",
  "name": "Knowledge Base for 06d99876-8219-4e74-84b7-04895692da59",
  "provider": "trieve",
  "createdAt": "2025-06-16T02:48:51.141Z",
  "updatedAt": "2025-07-29T04:23:56.309Z",
  "vectorStoreProviderId": "2c22683e-35cd-4fec-80f0-37f50264a88f",
  "vectorStoreSearchPlan": {
    "searchType": "fulltext",
    "scoreThreshold": 0,
    "removeStopWords": true
  },
  "vectorStoreCreatePlan": {
    "fileIds": [
      "19f4ebc0-3e24-4407-9418-d273736e269c"
    ]
  }
}

Thanks @Shubham Bajaj @ivy https://cdn.discordapp.com/attachments/1399683563187601408/1399683564001427556/message.txt?ex=6889e466&is=688892e6&hm=b565426346647fcb6a5144d351380da369ecd484fa580f75f33b454d3ad6d057&

Hi Maybe you have the same problem like me https://discord.com/channels/1211482211119796234/1392799267470442587/1392799267470442587

Thank you, i appreciate your help

Vapi no longer supports fileIds for Trieve instead, you’ll create a dataset in Trieve, then import it into Vapi using its dataset ID via the createPlan: { type: "import", providerId: "<dataset_id>" } format, and attach that knowledge base to your assistant..

I fixed the issues with change the model to gpt-4o or stronger model. Hope this is works for others

Glad it fixed