Azure AI services

Azure speech containers testing

If the speech container is WebSocket-based and REST APIs are not supported, how does Swagger work? Swagger doesn't seem to work with the Speech-to-Text container. Please provide ways to verify the capabilities of the speech container. I am pulling the…

I am not able to add my Azure AI Model on cursor

ETA for OpenAI Service quotas raise?

Hello all, Does Microsoft have an ETA for the quota increase requests now that they are managed through a form?? Thank you in advance. Src: https://learn.microsoft.com/en-us/azure/ai-services/openai/quotas-limits#how-to-request-quota-increases

Azure OpenAI Phi-4-multimodel-instruct: 'auto' tool choice error when using runTools() method that worked with GPT-4o

I recently switched from using GPT-4o to Phi-4-multimodel-instruct in my Next.js application using Azure AI services, but I'm encountering the following error: BadRequestError: 400…

Using PTU vs pay as you go in production LLMs models

Hi everyone I have seen in a Microsoft video that pay as you go sku models are more for labs and dev purposes and not suited for production workloads, they advice to use PTU sku models in production instead, to guarantee stable throughput. My question…

Connection Default_AzureOpenAI can't be found in this workspace.

I am trying to deploy gpt-4o model in Azure AI Foundry. After I created a new project and tried to deploy, I get the following error. Failed to get the connection NotFoundError: Connection Default_AzureOpenAI can't be found in this workspace. Please…

Is it possible for a model to return a hyperlink or URL in his response?

Hi, I created a simple chatbot using prompt flow in AI Foundry. I was wondering if it is possible for the model to answer with hyperlink or a URL. I want the chatbot to include a hyperlink to the public file stored on our web page based on which he…

The specified scale type 'Standard' of account deployment is not supported by the model 'gpt-4o

I am trying to deploy a Base model via rest api but i am passing the following error. Failed to deploy model: 400, {"error":{"code":"InvalidResourceProperties","message":"The specified scale type 'Standard'…

What quota do I need increased to deploy serverless models in AI Foundry

I am trying to deploy the Ministral-3B model from the Azure AI foundry model catalog as a serverless API. But when trying to do so I am getting the message shown in the image: I click on Manage Quota: I then click on Request Quota screen but this…

where all am i wrong??

Azure Ai Services(used) Ai search Creation : service/resource is created under the Azure -> Azure Ai Services -> AI Search.  For searching through the DataBase, Vector database. An embedding model needs to be deployed in order to use…

Why Azure AI foundry HUB comes by default with ai services resource

Hello We are planning a RAG system with multi models deployments so we plan to use AI foundry. When creating an AI foundry HUB an Azure ai services resource is created by default and is mandatory. Why this resource is mandatory to create for the HUB ?

Grounding with Bing Search is not working

According to this article: https://azure.microsoft.com/en-us/updates?id=483570 we need to replace Bing Search v7.0 with Grounding With Bing Search in "Agents" as described here:…

Deploying Deepseek R1 as Serverless API with .models.ai.azure.com endpoint format

I have a problem with creating a deployment of Deepseek R1 of "Serverless" as opposed to "Azure AI Services" type. I need the endpoint in the format of [deploymentname.eastus2.models.ai.azure.com] NOT…

Phi4 multimodal tools and function calling

Does Phi4 multimodal support tools and function calling?

How do I evaluate reasoning models (o3-mini) using Azure AI Foundry evaluation or Prompt Flow?

I have tried to evaluate the o3-mini model using an existing dataset, via Azure Machine Learning Workspace PromptFlow, as well as in Azure AI Foundry Evaluation, Azure OpenAI Evaluation, and PromptFlow options. However, it failed due to the parameter…

Unable to select the DeepSeek or any other opensource model deployments when creating the agent in Azure Foundry

I have models deployed for OpenAI gpt 4 and also Deep Seek R1, however when creating the agent in Azure AI Foundry I'm not able to select the other open source models that are already created as deployment. The drop down doesn't provide the…

Why am I getting rate limited?

My deployment's limits are 8k Tokens per minute 48 Requests per minute I'm using GPT-4o + OpenAI sdk (Assistants). Every other streaming request is getting rate limited -- LastError(code='rate_limit_exceeded', message='Rate limit is exceeded. Try…

Performance Issue with Azure DeepSeek Model – High Response Time

We have created and deployed the Azure DeepSeek model following the guidelines provided in this documentation link. While the model is functional, we are experiencing significant latency issues. A single query takes more than one minute to generate a…

Document intelligence service location is different that was chosen

I created a new Document Intelligence service in a resource group located in West Europe. The DI service is also located in West Europe. However, when I check the cost grouped by resources, I see that all costs associated with this service come from the…

Azure AI & Serverless functions for testing purpose.

Isn't there any test environment to test azure AI with serverless functions? I want to test my application before I purchase any compute or storage service. I really need this. I have developed a very nice service based on azure ai using serverless…