azure-ai-projects-py

Manage Azure Foundry AI agents and project resources with the azure-ai-projects Python SDK.

Updated Apr 12, 2026
One-click install
npx skills add https://github.com/davidrrowley/CortexYouV3 --skill azure-ai-projects-py-davidrrowley
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: azure-ai-projects-py
Source: https://github.com/davidrrowley/CortexYouV3/tree/main/.agents/skills/azure-ai-projects-py
Command: npx skills add https://github.com/davidrrowley/CortexYouV3 --skill azure-ai-projects-py-davidrrowley

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires azure-ai-projects, azure-identity, openai, and includes scripts (resource) and references (resource) components.

What problem does it solve?

Manage and orchestrate Azure Foundry project-level AI resources programmatically to avoid manual, error-prone workflows and speed agent development and evaluation.

Core Features & Use Cases

  • Agent lifecycle management: Create, version, list, and delete PromptAgentDefinition and hosted agents.
  • Evaluations and grading: Run batch and built-in/custom evaluations using an OpenAI-compatible client and cloud evals.
  • Project integrations: Manage connections, deployments, datasets, indexes, memory stores, and tool integrations (Code Interpreter, File Search, Azure AI Search, Bing).
  • Use Case: Create a versioned customer-support agent that uses file search and function tools, run evaluation suites against recorded conversations, and iterate on versions based on evaluator metrics.

Quick Start

Create a project client, list available deployments, and create a versioned PromptAgentDefinition named "support-agent" for immediate testing.

Frequently Asked Questions about azure-ai-projects-py

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I manage Azure Foundry agents programmatically using Python?

You can manage Azure Foundry agents programmatically using the azure-ai-projects Python SDK to create, version, list, and delete PromptAgentDefinition and hosted agents. Use the AIProjectClient to handle both synchronous and asynchronous operations for your project workflows.

Does the azure-ai-projects SDK support running batch evaluations against recorded conversations?

Yes, the azure-ai-projects SDK supports running batch and built-in or custom evaluations against recorded conversations. It uses an OpenAI-compatible evaluation client and cloud evals to help you iterate on agent versions based on evaluator metrics.

Can I use Azure AI Search and Bing grounding with Azure Foundry agents?

Yes, Azure Foundry agents built with the azure-ai-projects SDK support tool integrations including Azure AI Search and Bing grounding. You can also use CodeInterpreterTool and FileSearchTool to enhance your agent's capabilities.

What's the best way to create a versioned customer-support agent in Azure Foundry?

The best way to create a versioned customer-support agent is using the azure-ai-projects SDK to define a PromptAgentDefinition. You can integrate file search and function tools, then run evaluation suites to iterate based on evaluator metrics.

Do I need azure-identity and openai packages to build Foundry AI apps?

Yes, you need azure-identity and openai packages as dependencies to build Foundry AI apps. The azure-ai-projects SDK relies on azure-identity for authentication and the openai package for evaluation client compatibility.

Can I manage datasets and indexes within Azure Foundry project workflows?

Yes, you can manage datasets and indexes within Azure Foundry project workflows using the AIProjectClient. The SDK also supports managing connections, deployments, and memory stores to orchestrate project-level AI resources programmatically.