MCP Server that interacts with Azure AI Foundry (experimental)
A Model Context Protocol server for Azure AI Foundry, providing a unified set of tools for models, knowledge, evaluation, and more.
接入你的智能体
從發布者安裝。不是 PlugMCP 端點。
遠端 URL 還是本機指令? 兩種安裝方式的差異
MCP Server for Azure AI Foundry (experimental) has moved to the cloud, now as Foundry MCP Server (preview). Please check the official public documentation at Get started with Foundry MCP Server (preview). This GitHub repository contains the outdated information on the previous experimental server implementation, and will not be updated further. Instead, refer to the new, remotely hosted and managed Foundry MCP Server, that has the following benefits:
The rest of this repo (including codebase) explains the now-deprecated MCP Server that interactes with Azure AI Foundry (experimental).
This GitHub template has minimal setup with MCP server configuration and all required dependencies, making it easy to get started with your own projects.
This helps you automatically set up the MCP server in your VS Code environment under user settings. You will need
uvxinstalled in your environment to run the server.
Install uv by following Installing uv.
Start a new workspace in VS Code.
(Optional) Create .env file in the root of your workspace to set environment variables.
Create .vscode/mcp.json in the root of your workspace.
{
"servers": {
"mcp_foundry_server": {
"type": "stdio",
"command": "uvx",
"args": [
"--prerelease=allow",
"--from",
"git+https://github.com/azure-ai-foundry/mcp-foundry.git",
"run-azure-ai-foundry-mcp",
"--envFile",
"${workspaceFolder}/.env"
]
}
}
}
Click Start button for the server in .vscode/mcp.json file.
Open GitHub Copilot chat in Agent mode and start asking questions.
See More examples for advanced setup for more details on how to set up the MCP server.
To securely pass information to the MCP server, such as API keys, endpoints, and other sensitive data, you can use environment variables. This is especially important for tools that require authentication or access to external services.
You can set these environment variables in a .env file in the root of your project. You can pass the location of .env file when setting up MCP Server, and the server will automatically load these variables when it starts.
See example .env file for a sample configuration.
[!NOTE] Model
GITHUB_TOKENis used to authenticate with GitHub API for testing models. It is not required if you are exploring models from Foundry catalog.Knowledge
- See Create a search service to learn more about provisioning a search service.
- Azure AI Search supports multiple authentication methods. You can use either a Microsoft Entra authentication or an Key-based authentication to authenticate your requests. The choice of authentication method depends on your security requirements and the Azure environment you are working in.
- See Authenication to learn more about authentication methods for a search service.
Evaluation
- If you're using agent tools or safety evaluators, make sure the Azure project credentials are valid.
- If you're only doing text quality evaluation, the OpenAI endpoint and key are sufficient.
MIT License. See LICENSE for details.
A MCP Server for Azure AI Foundry: it's now moved to cloud, check the new Foundry MCP Server
工具
尚無工具清單。遠端伺服器會在首次健康檢查後取得工具清單。
常見問題
在本頁複製適用於 Cursor、Claude Desktop、Claude Code 或 VS Code 的 stdio 設定。它使用發布者的安裝方式,而不是 PlugMCP 代理伺服器。
這筆記錄未列出必要的環境變數。
MCP Foundry 透過 stdio 在本機執行。
上次檢查時間:2026-10-28。 A MCP Server for Azure AI Foundry: it's now moved to cloud, check the new Foundry MCP Server