发现优秀的 MCP 服务器

Vivado MCP Native

Enables AI clients to directly launch, control, and analyze AMD/Xilinx Vivado on Windows and Linux, supporting project management, synthesis, implementation, bitstream generation, timing/resource analysis, and simulation through a persistent Tcl session.

MCP-Coinglass

Provides futures market analytics (Open Interest, Liquidations, Long/Short Ratio, Funding Rates) via the CoinGlass API v4, enabling natural language queries for crypto derivatives data.

ClickHouse Cloud API MCP Server

A Multi-Agent Conversation Protocol server that enables interaction with the ClickHouse Cloud API, providing programmatic access to ClickHouse Cloud services through natural language.

ClinicalTrials.gov MCP Server

Empowers AI agents with direct access to the official ClinicalTrials.gov database, enabling programmatic searching, retrieval, and analysis of clinical study data through a Model Context Protocol interface.

postlayer-connect

A local MCP server that enables AI assistants to post to social media platforms (LinkedIn, X, Bluesky, Mastodon, Reddit, etc.) using your own app credentials, with encrypted token storage and no middleman.

MCP Server

A testing environment for Model Context Protocol that enables exploration of MCP capabilities and integration of AI models with external data sources and tools.

ourgroceries-mcp

Manage grocery lists on OurGroceries.com via CLI or MCP, supporting item operations, list management, and natural-language resolution.

Puppeteer

MCP Puppeteer 服务器 HomeAssistant 插件

ms-graph-mcp

Exposes Azure Entra ID user and license management as MCP tools over HTTP-SSE, enabling operations such as user creation, group assignment, and license management via Microsoft Graph API.

Mcpmapserver

为了学习，谷歌地图的 MCP 服务器。 (For study, Google Maps MCP server.) A more natural translation, depending on the context, could be: * **用于学习的谷歌地图 MCP 服务器:** (Google Maps MCP server for learning) - This is a more direct and common way to phrase it. * **研究谷歌地图 MCP 服务器:** (Researching Google Maps MCP server) - If the intention is to study the server itself. "MCP" likely refers to "Map Control Protocol" or something similar within the Google Maps architecture.

perplexity-mcp

Enables web search using Perplexity AI's API, allowing users to search the web with optional recency filters and integration with Claude, Cursor, and other MCP clients.

Everything Search MCP Server

使用特定于平台的技术，在 Windows、macOS 和 Linux 操作系统上提供快速的文件搜索功能。

temporal-mcp

Provides LLM agents with a sense of time between turns via two MCP tools that track elapsed time and day rollover per conversation thread.

persistenceone-bridgekitty

Cross-chain bridge aggregator MCP server for AI agents. Compares routes across LI.FI, deBridge, Relay, Across and Squid to find the best rate. Use when an agent needs to bridge or swap tokens between EVM chains, Solana, or Cosmos. The aggregator of aggregators.

Azure Model Context Protocol (MCP) Hub

Okay, here's a breakdown of resources for building and integrating Model Context Protocol (MCP) servers on Azure using multiple languages. Since MCP is a relatively new and evolving area, direct, comprehensive "one-stop-shop" resources are still emerging. I'll provide the best available information, focusing on the core components and how to adapt them to different languages on Azure. **Understanding Model Context Protocol (MCP)** * **Core Concept:** MCP is designed to provide a standardized way for AI models to access contextual information (e.g., user data, environment data, session history) at runtime. This allows models to make more informed and personalized decisions. * **Key Components:** * **MCP Server:** The central component that manages and serves the contextual data. This is what you'll be building. * **MCP Client:** The code within your AI model or application that requests data from the MCP Server. * **Data Sources:** The systems that hold the contextual information (databases, APIs, caches, etc.). **General Approach for Building an MCP Server on Azure** 1. **Choose a Language/Framework:** Select a language and framework suitable for building a web API. Popular choices include: * **Python (with Flask or FastAPI):** Excellent for rapid development and has a large ecosystem of libraries. * **C# (.NET):** Strong performance, well-suited for enterprise applications, and integrates seamlessly with Azure services. * **Node.js (with Express):** Good for building scalable and real-time applications. * **Java (with Spring Boot):** Another robust option for enterprise-grade solutions. 2. **Define the MCP API:** Design the API endpoints that your MCP Server will expose. This will likely involve: * **Request Format:** How the client will request data (e.g., using a specific ID or set of parameters). JSON is a common choice. * **Response Format:** The structure of the data returned by the server (again, likely JSON). * **Authentication/Authorization:** How you'll secure the API to ensure only authorized clients can access the data. 3. **Implement the API Logic:** Write the code to: * Receive requests. * Fetch data from the appropriate data sources. * Transform the data into the required response format. * Handle errors gracefully. 4. **Deploy to Azure:** Choose an Azure service to host your MCP Server: * **Azure App Service:** A fully managed platform for hosting web applications. Good for most scenarios. * **Azure Functions:** Serverless compute, ideal for event-driven architectures or APIs with infrequent usage. * **Azure Kubernetes Service (AKS):** For more complex deployments requiring container orchestration. * **Azure Container Apps:** A serverless container service that simplifies deploying containerized applications. 5. **Secure the API:** Implement authentication and authorization. Options include: * **Azure Active Directory (Azure AD):** For enterprise identity management. * **API Keys:** A simpler approach for less sensitive data. * **Managed Identities:** Allow your Azure resources to authenticate to other Azure services without needing to manage credentials. 6. **Monitor and Log:** Use Azure Monitor to track the performance and health of your MCP Server. Implement logging to help diagnose issues. **Language-Specific Resources and Examples (Adaptable for MCP)** While direct MCP examples in multiple languages are scarce, you can adapt existing Azure API examples: * **Python (Flask/FastAPI):** * **Azure App Service with Python:** [https://learn.microsoft.com/en-us/azure/app-service/quickstart-python](https://learn.microsoft.com/en-us/azure/app-service/quickstart-python) * **Azure Functions with Python:** [https://learn.microsoft.com/en-us/azure/azure-functions/functions-create-first-function-python](https://learn.microsoft.com/en-us/azure/azure-functions/functions-create-first-function-python) * **Example (Conceptual):** You would adapt these examples to: * Define API endpoints for your MCP data requests (e.g., `/context/{user_id}`). * Fetch data from your data sources (e.g., Azure Cosmos DB, Azure SQL Database). * Return the data in a JSON format. * **C# (.NET):** * **Azure App Service with .NET:** [https://learn.microsoft.com/en-us/azure/app-service/quickstart-dotnetcore](https://learn.microsoft.com/en-us/azure/app-service/quickstart-dotnetcore) * **Azure Functions with C#:** [https://learn.microsoft.com/en-us/azure/azure-functions/functions-create-first-function-vs](https://learn.microsoft.com/en-us/azure/azure-functions/functions-create-first-function-vs) * **Example (Conceptual):** Similar to Python, you'd create API controllers to handle MCP requests, access data sources using Entity Framework or other data access libraries, and return JSON responses. * **Node.js (Express):** * **Azure App Service with Node.js:** [https://learn.microsoft.com/en-us/azure/app-service/quickstart-nodejs](https://learn.microsoft.com/en-us/azure/app-service/quickstart-nodejs) * **Azure Functions with Node.js:** [https://learn.microsoft.com/en-us/azure/azure-functions/functions-create-first-function-node](https://learn.microsoft.com/en-us/azure/azure-functions/functions-create-first-function-node) * **Example (Conceptual):** Use Express to define routes for your MCP API, connect to data sources using libraries like `pg` (for PostgreSQL) or `mongodb` (for MongoDB), and return JSON data. * **Java (Spring Boot):** * **Azure App Service with Java:** [https://learn.microsoft.com/en-us/azure/app-service/quickstart-java](https://learn.microsoft.com/en-us/azure/app-service/quickstart-java) * **Azure Functions with Java:** [https://learn.microsoft.com/en-us/azure/azure-functions/functions-create-first-java](https://learn.microsoft.com/en-us/azure/azure-functions/functions-create-first-java) * **Example (Conceptual):** Use Spring Boot's REST controller features to create API endpoints, use Spring Data JPA or JDBC to access databases, and return JSON responses. **Key Considerations for MCP Implementation** * **Data Consistency:** Ensure that the data served by your MCP Server is consistent and up-to-date. Consider using caching mechanisms (e.g., Azure Cache for Redis) to improve performance and reduce load on your data sources. * **Scalability:** Design your MCP Server to handle a large number of requests. Azure App Service and Azure Functions can scale automatically. For more demanding workloads, consider AKS. * **Latency:** Minimize the latency of data retrieval. Optimize your data queries and use caching effectively. Consider the geographic location of your MCP Server and your AI models. * **Security:** Protect the data served by your MCP Server. Use strong authentication and authorization mechanisms. Encrypt data in transit and at rest. * **Data Governance:** Implement policies to ensure that data is used responsibly and ethically. Comply with relevant data privacy regulations (e.g., GDPR, CCPA). **Example Scenario (Python with FastAPI on Azure App Service)** 1. **FastAPI App:** ```python from fastapi import FastAPI, HTTPException from azure.cosmos import CosmosClient, PartitionKey from typing import Optional import os app = FastAPI() # Azure Cosmos DB Configuration (replace with your actual values) COSMOS_ENDPOINT = os.environ["COSMOS_ENDPOINT"] COSMOS_KEY = os.environ["COSMOS_KEY"] DATABASE_NAME = "mcp_db" CONTAINER_NAME = "user_context" # Initialize Cosmos DB client cosmos_client = CosmosClient(COSMOS_ENDPOINT, COSMOS_KEY) database = cosmos_client.get_database_client(DATABASE_NAME) container = database.get_container_client(CONTAINER_NAME) @app.get("/context/{user_id}") async def get_user_context(user_id: str): """ Retrieves user context data from Cosmos DB. """ try: item = container.read_item(item=user_id, partition_key=user_id) return item except Exception as e: raise HTTPException(status_code=404, detail="User context not found") @app.get("/health") async def health_check(): return {"status": "ok"} ``` 2. **Deployment to Azure App Service:** * Create an Azure App Service instance. * Configure environment variables for `COSMOS_ENDPOINT` and `COSMOS_KEY`. * Deploy the Python code to the App Service. You'll likely need a `requirements.txt` file listing dependencies (e.g., `fastapi`, `azure-cosmos`). **Important Notes:** * **MCP is Evolving:** The Model Context Protocol is still under development. Expect changes and updates to the specifications and available tools. * **Customization is Key:** You'll need to tailor your MCP Server to the specific needs of your AI models and data sources. * **Security Best Practices:** Always prioritize security when building and deploying your MCP Server. I hope this comprehensive guide helps you get started with building and integrating MCP servers on Azure using multiple languages! Remember to adapt the examples and resources to your specific requirements.

artl-mcp

Enables comprehensive scientific literature retrieval and analysis through Europe PMC, PubMed, and other databases, supporting metadata extraction, full-text access, and identifier conversion via MCP and CLI.

MCP X++ Server

An MCP server for Microsoft Dynamics 365 Finance & Operations that enables the creation, modification, and analysis of D365 objects like classes, tables, and forms. It integrates with Visual Studio 2022 to provide tools for X++ code extraction, codebase search, and safe object deletion with dependency validation.

mcp-edd

MCP server for Easy Digital Downloads REST API, enabling access to sales data, customers, products, and analytics from your EDD store.

zotero-cli-agent

A lightweight, context-efficient CLI and optional stdio MCP server for semantic search, browsing, and writing to your Zotero library.

果子狸MCP服务

Connects Jia Li Chuang EDA / EasyEDA Pro schematics to MCP clients, enabling AI-driven schematic editing, reading, and analysis.

mcp-sofa

MCP server for Stack Overflow for Agents, enabling search, post creation, voting, and knowledge exchange for AI agents.

sre-toolkit-mcp

Enables LLM agents to perform SRE reliability calculations like error budgets and burn rates using deterministic tools, integrating with Prometheus and Loki for real data.

openrouter-mcp

Enables AI assistants to consult other AI models via OpenRouter for help with coding tasks.

mockscreenshots-mcp

Compose fake chat conversations and receive deep links to a pre-filled generator for preview, parody, and design mockups.

travel-planner

Enables travel planning through natural language, providing weather forecasts, attraction search, itinerary generation, and distance calculations using free APIs.

smart-bug-triage

MCP server that triages GitHub issues by classifying, ranking priority, and assigning an engineer using a local LLM (Ollama) with full traceability.

MySQL MCP Server Pro

Provides comprehensive MySQL database operations including CRUD, performance optimization, health analysis, and anomaly detection. Supports multiple connection modes, OAuth2.0 authentication, and role-based permissions for database management through natural language.

AI List My Business

Country-agnostic MCP-callable directory for AI agents to find local SMBs — realtors, insurance agents, medical practitioners — by category, location, or natural-language query. Returns business catalog data and UTM-tagged booking URLs (zero PII).

MCP Hub

An Express server implementation of Model Context Protocol that allows websites to connect to LLMs through streamable HTTP and stdio transports, with a built-in chat UI for testing responses.