Please advise Azure AI Deployment Error code: 400 'message': 'Invalid Azure OpenAI configuration detected: Invalid embedding endpoint or deployment'

Nihit Mody 50 Reputation points
2025-03-14T07:21:34.4666667+00:00

Hi,

I followed the tutorial https://learn.microsoft.com/en-us/azure/ai-foundry/tutorials/deploy-chat-web-app exactly. I am able to interface with my chat app in Azure AI Foundry Playground, when deploying as a web app however no matter what pricing plan I choose I receive this error when I enter text related to the data I added (writing just Hello returns a general response):

Error code: 400 - {'error': {'requestid': 'xxx-xxx-xxx-xxx-xx'', 'code': 400, 'message': 'Invalid Azure OpenAI configuration detected: Invalid embedding endpoint or deployment'}}

I have tried everything from rebuilding from scratch and deleting all resources multiple times and I am able to interact perfectly fine in the Azure AI Playground. I resolved the Manage Identity error by popularing AZURE_SEARCH_KEY for the web app. Could someone please tell me how to resolve this error?

Thanks

Steps to Reproduce: Follow this tutorial: https://learn.microsoft.com/en-us/azure/ai-foundry/tutorials/deploy-chat-web-app from start to finish verbatim gpt-4o-mini (version:2024-07-18) Upload file as .xls Vector+Keyword Works fine in Azure AI Playground

Expected behavior Chat returns like it does in Azure AI Playground.

gpt-4o-mini (version:2024-07-18)

Cchat history enabled? No

Data source? Upload file -> .xls

Azure AI services
Azure AI services
A group of Azure services, SDKs, and APIs designed to make apps more intelligent, engaging, and discoverable.
3,379 questions
{count} votes

Accepted answer
  1. JAYA SHANKAR G S 2,180 Reputation points Microsoft External Staff
    2025-03-25T09:22:35.53+00:00

    Hi @Nihit Mody ,

    How does this factor in, is the web app only usable safely through the REST API? I am concerned about reliability.

    Actually, Web app is needed while you want web application integrated with Azure OpenAI, REST api works once you deploy the model.

    After the discussion, I understand your requirement is that need to create a vector search index on your chunked data, also I can see you raised new case for that.

    Please accept and give feedback by clicking on Yes.

    Thank you

    1 person found this answer helpful.
    0 comments No comments

0 additional answers

Sort by: Most helpful

Your answer

Answers can be marked as Accepted Answers by the question author, which helps users to know the answer solved the author's problem.