发现优秀的 MCP 服务器

Chromate MCP

A Model Context Protocol server for Chrome tab automation using the Chrome DevTools Protocol, enabling AI agents to select tabs, capture screenshots, and interact with web pages via CSS viewport coordinates.

aemo-mcp

MCP server for the Australian Energy Market Operator (AEMO) National Electricity Market. Plain-English access to 5-min dispatch prices, regional demand, interconnector flows, generation by fuel, rooftop PV.

MCP SERVERLY

这是一个旨在表达在官方 Anthropic base SDK for dotnet 9 中模型上下文协议 (MCP) 实现的项目。

Breadcrumb MCP Server

Enables AI assistants to interact with the Breadcrumbs project memory system, allowing creation and management of project documentation including sessions, components, ADRs, patterns, and providing search capabilities across project knowledge.

Sophos Central MCP Server

Enables AI-powered security operations through natural language, managing endpoint security, email threats, firewall policy, and more across multiple Sophos tenants with 334 tools, designed for MSP/MSSP teams.

CodeBrain MCP Server

Enables semantic code search across codebases using AI embeddings and vector similarity, integrated with Claude Desktop and Cursor.

reddit-mcp

Provides access to public Reddit data through tools for searching subreddits, viewing posts, and reading comments without requiring authentication. It enables Model Context Protocol clients to interact with Reddit's public JSON API via stdio or HTTP transports.

Topaz MCP Server

Create Topaz AI tasks through RunAPI — 2 endpoints, 2 models — then poll status and check pricing.

SimBrief Flight Planning MCP Server

Enables access to SimBrief flight planning data through Claude Desktop with secure Google OAuth authentication. Provides tools to retrieve flight plans, dispatch briefings, NOTAMs, weather data, and other aviation planning information.

Tennis Warehouse MCP Server

Provides secure access to Tennis Warehouse product data via natural language queries, enabling product search, availability checks, and deal discovery.

Better Fetch

A Model Context Protocol server that intelligently fetches and processes web content, transforming websites and documentation into clean, structured markdown with nested URL crawling capabilities.

FS_MCP_thin

A read-only MCP server that connects AI assistants to Freshservice Service Desk for ticket triage, asset analysis, and support history summarization.

GenPilot

GenPilot 简化了由生成式人工智能驱动的多代理系统的创建。它遵循 MCP 协议，并通过直观的终端或 Web 界面确保与 MCP 服务器的顺利集成。

first-mcp-kit

A real, runnable MCP server exposing three read-only tools (inventory, pricing, hours) with a CLI agent that lets Claude discover and call them.

Perplexity MCP Server

Okay, I understand. You want to perform a web search using Perplexity to get information for an MCP (presumably a topic you'll provide later), and you want to do this *without* using any API keys. Here's the breakdown of why that's tricky and what options you *might* have, along with their limitations: **Why It's Difficult (and why API keys exist):** * **Perplexity's Design:** Perplexity is designed to be accessed primarily through its official website or its API. The API is the intended way for developers to programmatically access its search and summarization capabilities. APIs are protected by API keys to: * **Control Usage:** Prevent abuse and ensure fair access to resources. * **Track Usage:** Monitor how the service is being used for billing and performance analysis. * **Enforce Terms of Service:** Ensure users adhere to the rules of the platform. * **Web Scraping Challenges:** The alternative to using an API is web scraping (programmatically extracting data from a website). This is generally discouraged and can be problematic: * **Terms of Service Violation:** Most websites (including Perplexity) explicitly prohibit scraping in their terms of service. Violating these terms can lead to your IP address being blocked. * **Website Structure Changes:** Websites change their HTML structure frequently. A scraper that works today might break tomorrow. * **Rate Limiting:** Websites often implement rate limiting to prevent scrapers from overloading their servers. You'll likely be blocked if you make too many requests too quickly. * **Ethical Considerations:** Scraping can put a strain on the website's resources and potentially disrupt service for other users. **Possible (But Limited and Potentially Unreliable) Approaches:** 1. **Manual Search and Copy-Pasting:** * **How it works:** The simplest approach is to manually go to the Perplexity website (perplexity.ai), enter your MCP search query, and then copy and paste the results into your own document or application. * **Pros:** No code required, avoids violating terms of service. * **Cons:** Extremely tedious and time-consuming for anything beyond a few searches. Not automated. 2. **Very Basic Web Scraping (Use with Extreme Caution):** * **Disclaimer:** I strongly advise against this unless you understand the risks and are prepared to deal with potential blocking or legal issues. *Only use this for very small, infrequent, and non-commercial purposes.* * **How it *might* work (but likely won't for long):** * **Inspect the Perplexity Website:** Use your browser's developer tools (usually by pressing F12) to examine the HTML structure of the Perplexity search results page. Identify the HTML elements that contain the search results you want (e.g., `<div>` tags, `<p>` tags, etc.). * **Use a Web Scraping Library:** Use a Python library like `requests` to fetch the HTML content of the Perplexity search results page for your query. Then, use a library like `Beautiful Soup` to parse the HTML and extract the data you identified in the previous step. * **Example (Conceptual - Likely to Break):** ```python import requests from bs4 import BeautifulSoup query = "Your MCP Search Query Here" # Replace with your actual query url = f"https://www.perplexity.ai/search?q={query}" #This is a guess at the URL structure try: response = requests.get(url) response.raise_for_status() # Raise an exception for bad status codes (404, 500, etc.) soup = BeautifulSoup(response.content, "html.parser") # **THIS IS THE TRICKY PART - YOU NEED TO FIND THE RIGHT HTML ELEMENTS** # Example: Let's say the search results are in <div> tags with class "result-item" results = soup.find_all("div", class_="result-item") for result in results: # Extract the text from each result (adjust based on the actual HTML) text = result.text.strip() print(text) except requests.exceptions.RequestException as e: print(f"Error: {e}") except Exception as e: print(f"An unexpected error occurred: {e}") ``` * **Pros:** Potentially automates the search process (but very fragile). * **Cons:** * **High risk of being blocked.** * **Very likely to break due to website changes.** * **Potentially violates Perplexity's terms of service.** * **Requires programming knowledge.** * **No guarantee of accurate or complete results.** **Important Considerations:** * **Ethical Scraping:** If you absolutely must scrape, be respectful: * **Identify Yourself:** Set a `User-Agent` header in your `requests` call to identify your script and provide contact information. * **Rate Limiting:** Introduce delays between requests (e.g., using `time.sleep()`) to avoid overloading the server. * **Check `robots.txt`:** Examine the website's `robots.txt` file (e.g., `perplexity.ai/robots.txt`) to see if there are any specific rules about which pages you are allowed to crawl. * **Consider Alternatives:** Before resorting to scraping, explore other search engines that might offer more accessible APIs or data feeds (even if they aren't exactly Perplexity). **In summary, performing a Perplexity web search without an API key is highly discouraged and comes with significant risks and limitations. The manual approach is the safest, but the least efficient. Web scraping is technically possible, but ethically questionable and practically unreliable.** To give you a more specific answer, please tell me: 1. **What is the MCP you want to search for?** 2. **What is the *purpose* of this search?** (e.g., personal research, academic project, commercial application). This will help me understand the context and suggest more appropriate solutions. 3. **What is your level of programming experience?** Once I have this information, I can provide more tailored advice. However, I must reiterate that I cannot endorse or assist with any activity that violates a website's terms of service.

MCP Server Hub / Gateway

这个项目提供了一个中心网关来管理多个 MCP（模型上下文协议）服务器，从而避免了为每个 LLM 客户端（如 Cline、Cursor 等）配置和运行重复服务器进程的需求。 将您的 LLM 客户端连接到此项目提供的单个网关客户端端点，即可访问来自所有受管理的 MCP 服务器的工具。

XSearchLane

MCP server for realtime X (Twitter) search enabling keyword/semantic search, filters, date ranges, and citations for coding agents.

Agence

Governs and audits AI coding agents via command gating, cryptographic logging, and multi-agent orchestration, exposed as an MCP server.

DevPilot MCP

An AI-powered development assistant that enables LLMs to securely interact with local development tools including Git, filesystem, Docker, PostgreSQL, and GitHub via the Model Context Protocol.

MCP Skill Registry

A self-hostable MCP server that turns a folder of skills into callable tools via MCP and REST APIs.

MIDI File MCP

一个强大的 MCP 工具，用于解析和操作 MIDI 文件，它允许用户通过自然语言命令读取、分析和修改 MIDI 文件，支持读取文件信息、修改音轨、添加音符和设置速度等操作。

awesome-ai-native-cloudflare-app MCP Server

Enables AI agents to interact with the app via MCP tools and resources, supporting self-improvement through goal-setting and metrics tracking on Cloudflare Workers.

NexusAPI MCP Server

Enables AI agents to interact with NexusAPI's asynchronous generation API, including model discovery, schema inspection, task creation, and status retrieval.

Sherlog MCP

Provides persistent IPython shell sessions per conversation with DataFrame-centric architecture, enabling stateful data analysis, CLI tool execution, and integration of external MCP servers within the same workspace context.

Megahorn

A Go CLI and MCP server for cross-posting to Twitter, LinkedIn, and Reddit from your terminal or AI agents, with secure local credential storage.

Stargraph

Enables evidence-first analytics over GitHub starred repositories, including search, similarity comparison, explanation, and clustering via MCP tools.

firstcall-demo-mcp-server

A read-only MCP server generated by FirstCall that looks up public GitHub user profiles by username. It demonstrates FirstCall-generated MCP artifacts with inputSchema, outputSchema, structuredContent, and tool annotations.

iitkgp-erp-mcp

MCP server for IIT Kharagpur ERP that enables querying grades, CGPA, attendance, timetable, placements, and more through any AI agent.

@bounceprotect/mcp

Email validation and SMTP verification for Claude Desktop, Cursor, and Claude Code