How to Connect Azure Agents to Superinterface
Azure Agents provides a cloud-based platform for building AI agents with Azure's infrastructure. By connecting Azure AI Agents to Superinterface, you can leverage powerful capabilities like file search, code interpreter, and function calling.
This guide will walk you through how to connect Azure AI Agents to Superinterface and enable native tools.
Step 1: Set Up Azure AI Agents
To use Azure AI Agents, you need to complete the following setup in Azure:
1. Create an Azure AI Foundry Project
Log in to the Azure Portal
Create a new Azure AI Foundry Project
Note down your:
Subscription ID
Resource Group name
Project Name
2. Create a Service Principal and Set Permissions
You'll need a service principal (App Registration) with the proper permissions for authentication.
Required Setup:
Create an App Registration in Azure Active Directory
The service principal needs Contributor role scoped to your Azure AI Project resource
The service principal also needs Cognitive Services User role
How to set this up in Azure Portal:
Go to your Azure AI Project resource in Azure Portal
Navigate to Access control (IAM) in the left sidebar
Click Add → Add role assignment
Add the following two role assignments:
Role: "Contributor"
Assign access to: User, group, or service principal
Select: Your app registration / service principal
Role: "Cognitive Services User"
Assign access to: User, group, or service principal
Select: Your app registration / service principal
From your App Registration, note down:
Tenant ID (from Overview page)
Client ID / Application ID (from Overview page)
Client Secret (create one in Certificates & secrets if you don't have one)
3. Get Your Project Endpoint
Navigate to your Azure AI Foundry project in Azure AI Foundry
Go to Project settings or Overview
Find your project endpoint - it will look like:
https://your-project-name.cognitiveservices.azure.com/Copy this endpoint URL
Step 2: Add Credentials to Superinterface
Go to the Assistants tab in Superinterface and either select an existing assistant or click New Assistant
In the AI Provider section, select Azure AI Project
In the Thread Storage and Run Execution section, select Azure Agents
Enter your credentials from Step 1:
Project Endpoint: Your project endpoint (e.g.,
https://your-project.cognitiveservices.azure.com/)Tenant ID: From the service principal creation
Client ID: From the service principal creation (appId)
Client Secret: From the service principal creation (password)
Click Save to apply the settings
After a few seconds, your connection will be validated
Step 3: Enable Native Tools (Optional)
Azure AI Agents supports two native tools that can be enabled in Superinterface:
Code Interpreter
The code interpreter tool allows your agent to write and execute Python code in a sandboxed environment.
To enable in Superinterface:
In your assistant's configuration, scroll to the Tools section
Enable Code Interpreter
Click Save
The code interpreter will be automatically injected and work with your Azure AI Agent.
File Search (Knowledge)
The file search tool enables your agent to search through documents and retrieve relevant information.
Basic Setup (Superinterface only):
In your assistant's configuration, scroll to the Tools section
Enable File Search
Click Save
The file search tool will be automatically injected and work with your Azure AI Agent.
Advanced Setup (Azure AI Foundry):
If you also enable file search in Azure AI Foundry, you can use advanced features like Azure AI Search integration and custom search configurations:
Create a Vector Store:
Navigate to your project in Azure AI Foundry
Go to Knowledge > Vector Stores
Click Create Vector Store
Give it a name and configure expiration settings (optional)
Upload Files:
Select your vector store
Click Add Files to upload documents
Supported formats: PDF, DOCX, TXT, JSON, Markdown, HTML, and code files
Maximum file size: 512 MB per file
Maximum tokens: 5,000,000 tokens per file
Capacity: Up to 10,000 files per vector store
Attach to Your Agent:
Open your agent's configuration in Azure AI Foundry
Under Tools, enable File Search
Under Knowledge, select your vector store or configure Azure AI Search
Configure additional search settings like ranking parameters and filters
Note: When file search is enabled in both Superinterface and Azure AI Foundry, the Azure configuration takes precedence, allowing you to use advanced features like Azure AI Search for knowledge retrieval.
Step 4: Test Your Agent
Once your agent is connected and tools are enabled, test it through Superinterface:
Open the chat interface for your connected agent
For code interpreter: Ask it to perform calculations or data analysis
For file search: Ask questions about the documents you uploaded
Your Azure AI Agent is now ready to use with all its capabilities
Ready to Explore More?
Your agent is now set up, but you can further customize its capabilities with advanced configurations. Check out our Functions Guide to learn how to add custom tools to your agent.
Need Help?
If you need assistance with Azure AI Agents or have questions about integration, feel free to reach out! You can write us an email at [email protected], or book a call with us. We're here to help!