MCP 服务器
开放的 Model Context Protocol 服务器注册表。按类别、传输方式或使用场景,为 AI 智能体寻找合适的工具、资源和提示词。
部分描述为试点机器翻译内容,尚未经过人工审核。
服务器
177
工具
471
类别
11
贡献者
155
MCP server for Exa's neural search API. Provides AI agents with powerful web search capabilities using embeddings-based semantic search. Returns clean, parsed content from web pages with relevance scoring. Supports filtering by domain, date range, and content type for precise information retrieval from the internet.
MCP server for Firecrawl's web scraping and crawling API. Converts any website into clean, LLM-ready markdown or structured data. Supports single page scraping, multi-page crawling with depth control, sitemap-based extraction, and batch operations. Handles JavaScript-rendered pages, bypasses common anti-bot measures, and returns structured content suitable for RAG pipelines and knowledge base construction.
Official MCP server for the Perplexity API Platform. Provides AI agents with real-time web search, deep research, and advanced reasoning capabilities through Sonar models. Includes tools for quick web search, conversational Q&A with citations, comprehensive deep research reports, and complex analytical reasoning. Returns answers with source attribution and supports configurable timeouts for long research queries.
通过 Model Context Protocol 调用 Brave Search API,提供完整的搜索能力。AI 智能体可以执行网页、本地商家、图片、视频和新闻搜索,并生成 AI 摘要。支持 STDIO 和 HTTP 传输。
MCP server for the Reddit API. Lets AI agents search subreddits, fetch hot/new/top posts, read comment threads, and retrieve user activity. Useful for market and community research, sentiment monitoring, trend discovery, and summarizing discussions across communities.
MCP server for Hacker News. Lets AI agents fetch top, new, best, Ask HN, and Show HN stories, read comment threads, and look up items and users via the official Firebase API. Useful for tracking tech trends, summarizing discussions, and research into what the developer community is talking about.
Community MCP server that provides privacy-friendly web search through DuckDuckGo, plus fetching and parsing of web page content into clean text. Lets AI agents look up current information and retrieve source pages without an API key, making it a lightweight option for research and retrieval workflows. Not affiliated with DuckDuckGo.
Google's official MCP server for Google Analytics 4. Lets AI agents run reports against the GA4 Data API, discover dimensions and metrics, and explore account and property metadata with server-side aggregation and safe defaults. Runs locally (via pipx) and connects to MCP-compatible clients for conversational analysis of website traffic and user behavior.
Official MCP server backed by the Kagi API. Exposes Kagi's high-quality search and summarization tools to MCP-compatible clients, including web search and the Universal Summarizer for pages and videos. Useful for research workflows that want ad-free, privacy-respecting results with concise summaries. Requires a Kagi API key.
Skills 与 MCP 服务器
有什么区别?Skills定义“做什么”
Skill 将说明、示例提示词和推荐模型组织在一起,让智能体稳定地完成任务。Skills 增加的是知识,而不是新的连接。
MCP 服务器定义“如何连接”
MCP 服务器通过连接数据库、API 和文件等真实系统,为智能体增加新能力。MCP 提供连接和操作,而不是任务说明。
简单来说:当模型需要把某项任务做好时选择 Skill;当模型需要访问工具或系统时选择 MCP 服务器。两者可以组合使用,Skill 可以依赖 MCP 服务器提供的工具。
开发了 MCP 服务器?
将它提交到由社区维护的开源注册表。