Unable to use azure Grounding with bing search with azure ai agent in python, Getting Resource not found error.

Question

Unable to use azure Grounding with bing search with azure ai agent in python, Getting Resource not found error.

Srija Lingamgunta 0

I am having issues when trying to run this code:

https://github.com/Azure/azure-sdk-for-python/blob/main/sdk/ai/azure-ai-agents/samples/agents_tools/sample_agents_bing_grounding.py

I am getting an error 404 resource not found.

Traceback (most recent call last):
  File "C:\Users\Downloads\bingtest\binggrounding.py", line 30, in <module>
    agent = agents_client.create_agent(
            ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\Downloads\bingtest\myenv\Lib\site-packages\azure\core\tracing\decorator.py", line 119, in wrapper_use_tracer
    return func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\Downloads\bingtest\myenv\Lib\site-packages\azure\ai\agents\_patch.py", line 294, in create_agent
    new_agent = super().create_agent(
                ^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\Downloads\bingtest\myenv\Lib\site-packages\azure\core\tracing\decorator.py", line 119, in wrapper_use_tracer
    return func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\Downloads\bingtest\myenv\Lib\site-packages\azure\ai\agents\operations\_operations.py", line 5308, in create_agent
    map_error(status_code=response.status_code, response=response, error_map=error_map)
  File "C:\Users\Downloads\bingtest\myenv\Lib\site-packages\azure\core\exceptions.py", line 163, in map_error
    raise error
azure.core.exceptions.ResourceNotFoundError: (404) Resource not found
Code: 404
Message: Resource not found

I have AI foundry resource in my Azure and I created Grounding with bing search resource. I connected the Grounding with Bing search resource in connected resources tab of Azure AI foundry resource.i have used the below string for Azure Bing Connection Id in my environmental variables.

/subscriptions/<your_subscription_id>/resourceGroups/<your_resource_group>/providers/Microsoft.CognitiveServices/accounts/<ai_service_name>/projects/<foundry-project-name>/connections/<connection_name_created_above>

I have used "ai_service_name" same as resource name(?) Is that right?

This is my AI foundry resource. I used Azure AI inference endpoint as Project Endpoint.

User's image

Amira Bedhiafi 34,491 Volunteer Moderator

Hello Srija !

Thank you for posting on Microsoft Learn.

Here are the three things that usually trigger the ResourceNotFoundError (404) when you try to call agents_client.create_agent() with the Bing-grounding sample:

What the service is looking for	Most common pitfall	How to double-check
Project endpoint – the base URL of your AI Foundry project.	Copying the “Azure AI model inference endpoint” (it ends in /models) instead of the project endpoint.	In the Foundry portal: Project ▸ Overview ▸ Endpoints and keys ▸ “Project endpoint”. It looks likehttps://<project-name>.<region>.services.azure.ai.azure.com/ (no path after the slash). (learn.microsoft.com)
Project endpoint – the base URL of your AI Foundry project.	Copying the “Azure AI model inference endpoint” (it ends in `/models`) instead of the project endpoint.	In the Foundry portal: Project ▸ Overview ▸ Endpoints and keys ▸ “Project endpoint”. It looks like`https://<project-name>.<region>.services.azure.ai.azure.com/` (no path after the slash). ([learn.microsoft.com](https://learn.microsoft.com/en-us/azure/ai-services/agents/how-to/tools/bing-code-samples?pivots=python"Grounding with Bing Search code samples - Azure AI Foundry	Microsoft Learn"))
Model deployment name – the exact “Name” value from Models + endpoints (e.g. `gpt-4o-mini`).	Using the model family (`gpt-4o`) instead of the deployment name, or a name from a different project.	Foundry portal ▸ Models + endpoints ▸ column Name.
3. Bing connection ID – the resource-ID of the “Grounding with Bing Search” connection that is attached to the same Foundry project.	▸ Hand-crafting the ID string by following doc samples (easy to get one segment wrong). ▸ Mixing the two formats that are circulating in the docs ( `Microsoft.CognitiveServices/.../accounts/...` vs `Microsoft.MachineLearningServices/.../workspaces/...`).	Let the SDK give you the value so you can’t mistype it (see snippet below).Whichever format the service currently uses for your subscription is what you’ll get back from the SDK.

An example that cannot pick up the wrong IDs

import os
from azure.identity import DefaultAzureCredential
from azure.ai.projects import AIProjectClient
from azure.ai.agents.models import BingGroundingTool, MessageRole

project_endpoint = os.environ["PROJECT_ENDPOINT"]
client = AIProjectClient(
    endpoint=project_endpoint,
    credential=DefaultAzureCredential(),   
)

bing_connection_name = os.environ.get("BING_CONNECTION_NAME", "groundingwithbingsearch")
conn_id = client.connections.get(connection_name=bing_connection_name).id
print(f"✔ Bing connection ID: {conn_id}")

bing_tool = BingGroundingTool(connection_id=conn_id)

model_deployment = os.environ["MODEL_DEPLOYMENT_NAME"]          
with client:
    agents = client.agents
    agent = agents.create_agent(
        model=model_deployment,
        name="my-bing-grounded-agent",
        instructions="You are a helpful assistant.",
        tools=bing_tool.definitions,
    )
    print(f"✔ Agent created: {agent.id}")
    thread = agents.threads.create()
    agents.messages.create(thread_id=thread.id,
                           role=MessageRole.USER,
                           content="How does Wikipedia explain Euler's Identity?")
    run = agents.runs.create_and_process(thread_id=thread.id, agent_id=agent.id)
    print(f"Run status: {run.status}")
    
    agents.delete_agent(agent.id)

The endpoint is the project root (no /models), so the service can find the project metadata.

MODEL_DEPLOYMENT_NAME is whatever the portal shows, so the model ID exists.

The call to client.connections.get() retrieves the exact connection ID string that the service expects.

At the moment, the ID string the SDK returns is the Microsoft.MachineLearningServices/workspaces/.../connections/... form – even though some docs still show the Cognitive Services form ([github.com](https://github.com/azure/azure-sdk-for-python/issues/41486"What is needed as a connection ID for using AI Agents with BING GROUNDING? · Issue #41486 · Azure/azure-sdk-for-python · GitHub")).

Pavankumar Purilla 8,660 Reputation points Microsoft External Staff Moderator

2025-07-08T04:28:39.09+00:00

Hi Srija Lingamgunta,
Did you get any chance to check the response. Thank you!

Srija Lingamgunta 0

Hi @Amira Bedhiafi @Pavankumar Purilla

I created a project inside my AI Foundry resource as suggested.
User's image

Now, i do see Azure AI foundry project endpoint, I added the same endpoint in the code.

But, I am getting following error. I do have Azure AI User role assigned for my account.

azure.core.exceptions.ClientAuthenticationError: (PermissionDenied) The principal `658c07c3-e711-453c5e` lacks the required data action `Microsoft.CognitiveServices/accounts/AIServices/agents/write` to perform `POST /api/projects/{projectName}/assistants` operation. For instructions on granting the necessary permissions, see https://aka.ms/FoundryPermissions.
Code: PermissionDenied
Message: The principal `658c07c3-e711-453c5e` lacks the required data action `Microsoft.CognitiveServices/accounts/AIServices/agents/write` to perform `POST /api/projects/{projectName}/assistants` operation. For instructions on granting the necessary permissions, see https://aka.ms/FoundryPermissions.

I have required Azure AI user role for my account assigned. Do i need any other role required?
User's image

Pavankumar Purilla 8,660 Reputation points Microsoft External Staff Moderator

2025-07-09T04:58:05.77+00:00

Hi Srija Lingamgunta,
Based on the current role assignments, your account has been granted the Azure AI User and Contributor roles on the AI Foundry resource. However, these roles do not include the specific data action Microsoft.CognitiveServices/accounts/AIServices/agents/write, which is required to create agents using the Azure AI Agents SDK. To resolve the PermissionDenied error, you will need to be assigned the Azure AI Project Contributor role on the same AI Foundry resource. This role includes the necessary permissions to perform agent creation and related write operations within an AI Foundry project. Please ask your Azure administrator to assign the Azure AI Project Contributor role to your account at the scope of the AI Foundry resource. Once this role is added, your code should work as expected. You can refer to the official documentation for more details: https://aka.ms/FoundryPermissions.
Pavankumar Purilla 8,660 Reputation points Microsoft External Staff Moderator

2025-07-10T11:35:04.8766667+00:00

Hi Srija Lingamgunta,
Just following up to see if you had a chance to review the above response. Thank you!

Your answer

Pavankumar Purilla 8,660 Reputation points Microsoft External Staff Moderator

2025-07-08T04:28:39.09+00:00

Hi Srija Lingamgunta,
Did you get any chance to check the response. Thank you!
Srija Lingamgunta 0 Reputation points

2025-07-08T12:19:52.44+00:00

Hi @Amira Bedhiafi @Pavankumar Purilla

I created a project inside my AI Foundry resource as suggested.

Now, i do see Azure AI foundry project endpoint, I added the same endpoint in the code.

But, I am getting following error. I do have Azure AI User role assigned for my account.

azure.core.exceptions.ClientAuthenticationError: (PermissionDenied) The principal `658c07c3-e711-453c5e` lacks the required data action `Microsoft.CognitiveServices/accounts/AIServices/agents/write` to perform `POST /api/projects/{projectName}/assistants` operation. For instructions on granting the necessary permissions, see https://aka.ms/FoundryPermissions. Code: PermissionDenied Message: The principal `658c07c3-e711-453c5e` lacks the required data action `Microsoft.CognitiveServices/accounts/AIServices/agents/write` to perform `POST /api/projects/{projectName}/assistants` operation. For instructions on granting the necessary permissions, see https://aka.ms/FoundryPermissions.

I have required Azure AI user role for my account assigned. Do i need any other role required?
Pavankumar Purilla 8,660 Reputation points Microsoft External Staff Moderator

2025-07-09T04:58:05.77+00:00

Hi Srija Lingamgunta,
Based on the current role assignments, your account has been granted the Azure AI User and Contributor roles on the AI Foundry resource. However, these roles do not include the specific data action Microsoft.CognitiveServices/accounts/AIServices/agents/write, which is required to create agents using the Azure AI Agents SDK. To resolve the PermissionDenied error, you will need to be assigned the Azure AI Project Contributor role on the same AI Foundry resource. This role includes the necessary permissions to perform agent creation and related write operations within an AI Foundry project. Please ask your Azure administrator to assign the Azure AI Project Contributor role to your account at the scope of the AI Foundry resource. Once this role is added, your code should work as expected. You can refer to the official documentation for more details: https://aka.ms/FoundryPermissions.
Pavankumar Purilla 8,660 Reputation points Microsoft External Staff Moderator

2025-07-10T11:35:04.8766667+00:00

Hi Srija Lingamgunta,
Just following up to see if you had a chance to review the above response. Thank you!

Share via

Unable to use azure Grounding with bing search with azure ai agent in python, Getting Resource not found error.

Your answer