Skip to content

gbaeke/agent_service

Repository files navigation

Azure AI Agent Service Demo

This code is discussed in the following post: https://blog.baeke.info/2024/12/24/creating-an-agent-with-the-azure-ai-agent-sdk/

Ensure you have an Azure AI Foundry hub and project. The hub should have gpt-4o-mini deployed. You can use another model if you wish but then you have to change the Python code in agent_api.py.

Ensure you have the Azure CLI and use it to login to your Azure account. Use the command az login to login. Your account needs AI Developer role on the project.

In the main folder, create a .env file with the following content:

PROJECT_CONNECTION_STRING="connection string to Azure AI Foundry project"
JWT_SECRET="JWT secret if you want to protect routes"

If you do not want to protect routes, set JWT_SECRET to anything and check the routes to comment/uncomment the function signature.

In folder agentui, do the following:

  • Install Python requirements: pip install -r requirements.txt
  • Run the app: python agent_api.py
  • Use VS Code Live Server to run index.html in folder agentui

About

Azure AI Agent Service Samples

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published