Stophy

io.github.stophydotdev由 GitHub 账号 stophydotdev 发布v1.3.1更新于 Nov 7, 2026

Give your agents the live web.

已验证Streamable HTTP可网页运行网页搜索与抓取浏览器自动化

服务详情

标识符: io.github.stophydotdev/mcp
传输: Streamable HTTP
运行时: Web executable
信任: Verified
发布于: Sep 28, 2026

安全提示

MCP 服务会接收对话中的工具调用请求，并可能执行外部操作。启用前请先了解该服务能做什么，并只授予它所需的权限。

注册中心已验证此条目所属的 GitHub 账号 stophydotdev。这并不表明该账号就是产品方。

SourceWeft 的检查只涵盖条目是否完整、端点是否响应；不审查服务器的代码，也不审查其工具的行为。

在 README 中添加徽章

向读者展示该 MCP 服务已收录于 SourceWeft，徽章链接到本页。

[![Listed on SourceWeft](https://sourceweft.com/badge/mcp/io.github.stophydotdev%2Fmcp.svg)](https://sourceweft.com/mcp/io.github.stophydotdev%2Fmcp?utm_source=readme&utm_medium=badge)

概览

AI 生成的概览

Stophy 让助手获取实时网页数据：搜索、抓取、列出站点页面、爬取和抽取字段，并可调用付费数据端点，支持托管或本地 MCP 服务器。

可以做什么: 搜索网页、新闻、图片、学术论文、公司和人物
把页面读取为 Markdown、HTML、链接、截图或 JSON，并可用站点专用解析器取结构化数据
列出站点页面、在后台爬取网站并查询爬取状态
按指定字段从页面抽取 JSON
按 id 查找、描述并调用数据端点，每次调用显示费用
发起注册流程，返回待确认链接，随后取得 API 密钥
适用场景: 当助手需要当前网页内容而非模型记忆时使用：调研、监控页面、收集站点数据，或从公开页面抽取结构化字段。适合把搜索、阅读与抽取放在一处完成的工作流。
运行要求: 使用托管端点（浏览器登录或 Authorization bearer 令牌），或使用本地 npm 包 @stophy/mcp，后者需要 Node.js 18 或更高版本并把请求转发到托管服务器。可选 STOPHY_API_KEY 可解锁全部端点；不设置时只能用免密钥端点，限制为每 IP 每分钟 10 次、每天 50 次。需要能访问 Stophy 主机的网络。

安装前请注意

调用会消耗额度即产生费用：多数调用 1 额度，搜索每 10 条结果 2 额度，爬取每页 1 额度，抽取多数页面 5 额度，checkPromptInjection 等可选功能另加 4 额度。API 密钥是可计费的机密，通过 STOPHY_API_KEY 或 Authorization bearer 令牌传递。注册工具返回需由本人确认的领取链接。页面内容和网址会发送到 Stophy 服务。

安装

在 SourceWeft 中

打开控制台中的 Stophy，将其添加到工作区。
为需要使用其工具的对话启用该服务。

Web executable，通过 Streamable HTTP。远程服务在工作区中配置后即可从网页运行时运行。

交给智能体安装

把这段提示词粘贴到 Claude Code、Codex、Cursor 等智能体，或 SourceWeft 对话中。智能体会读取该服务器的安装指南，先向你展示发布者、端点和登录方式，经你同意后再把它添加到你的客户端。密钥不会写进任何命令。

阅读 https://sourceweft.com/mcp/io.github.stophydotdev%2Fmcp/install.md，并按其中说明把这个 MCP 服务器接入我的客户端。修改任何配置之前，先告诉我它的发布者、端点和登录方式，等我确认。不要把 API 密钥写进命令或我会提交的文件；问我密钥应从哪里获取。

查看智能体所依据的安装指南

其他 MCP 客户端

把它添加到你客户端的 mcpServers 配置中。

{
  "mcpServers": {
    "stophy": {
      "type": "http",
      "url": "https://api.stophy.dev/mcp"
    }
  }
}

README

Stophy MCP

Give your agents the live web over MCP. Your agent can search, scrape, map, crawl and extract, and read a site's own data through a parser.

[Add to Cursor]

Connect to the hosted server

Most apps connect to the hosted server directly, with nothing to install.

URL	How you connect	What your agent can use
`https://api.stophy.dev/mcp`	No key	The endpoints marked `keyless: true` in the endpoint list, within a free allowance
`https://api.stophy.dev/mcp`	`Authorization: Bearer <key>` header	Every endpoint
`https://mcp.stophy.dev/mcp`	Sign in with your browser	Every endpoint

Get an API key at stophy.dev/signup.

Claude

In Claude on the web, Desktop or mobile, open Customize > Connectors, choose Add custom connector, and paste https://mcp.stophy.dev/mcp. Sign in when Claude asks.

Claude Code

To sign in with your browser, add the server, then run /mcp and choose Authenticate:

To use an API key:

bash

claude mcp add --transport http stophy https://api.stophy.dev/mcp --header "Authorization: Bearer <key>"

claude mcp add --transport http stophy https://api.stophy.dev/mcp --header "Authorization: Bearer <key>"

To start without a key:

Codex

To use the key in your STOPHY_API_KEY environment variable:

To sign in with your browser instead, add the sign-in URL, then log in:

Cursor

Use the Add to Cursor button above to sign in with your browser, or add this to .cursor/mcp.json to use a key:

json

{  "mcpServers": {    "stophy": {      "url": "https://api.stophy.dev/mcp",      "headers": { "Authorization": "Bearer <key>" }    }  }}

{  "mcpServers": {    "stophy": {      "url": "https://api.stophy.dev/mcp",      "headers": { "Authorization": "Bearer <key>" }    }  }}

To start without a key, leave out headers.

Run it as a local server

Some apps can only start a local MCP server. For those, use this package. It runs on your computer and passes every request to the hosted server, so it always has the same tools.

With STOPHY_API_KEY set, your agent can use every endpoint. Without it, your agent can use the endpoints that need no key, within a free allowance.

Claude Desktop

Add this to claude_desktop_config.json, then restart Claude Desktop:

json

{  "mcpServers": {    "stophy": {      "command": "npx",      "args": ["-y", "@stophy/mcp"],      "env": { "STOPHY_API_KEY": "<key>" }    }  }}

{  "mcpServers": {    "stophy": {      "command": "npx",      "args": ["-y", "@stophy/mcp"],      "env": { "STOPHY_API_KEY": "<key>" }    }  }}

To start without a key, leave out env.

The package needs Node.js 18 or later.

Tools

Tool	What it does
`stophy_search`	Searches the web, news and images
`stophy_scrape`	Reads a page as markdown, HTML, links, a screenshot or JSON
`stophy_map`	Lists the pages on a site
`stophy_crawl`	Starts a crawl of a site in the background and returns its `id`
`stophy_extract`	Pulls the fields you name from pages as JSON
`stophy_search_endpoints`	Finds the endpoint for a task, with its cost. Free
`stophy_describe_endpoint`	Shows an endpoint's inputs, an example and what it returns. Free
`stophy_call`	Runs any endpoint by `id` with `input`, and optional `fields` and `maxTextLength`
`stophy_signup`	Gets your agent an API key: it returns a link for you to approve, then the key once you call it with the code

To get a site's own data as JSON, such as a Reddit search, a YouTube transcript or Google Maps places, call stophy_scrape with the site's page URL and a parser:

json

{  "url": "https://www.reddit.com/search/?q=bun+runtime",  "formats": [{ "type": "json", "parser": "reddit-search" }]}

{  "url": "https://www.reddit.com/search/?q=bun+runtime",  "formats": [{ "type": "json", "parser": "reddit-search" }]}

The stophy_scrape description names every parser, and the parser list shows the page each one takes.

Endpoint ids are camelCase, such as crawlStatus or batchScrape. A crawl runs in the background: call crawlStatus through stophy_call with the crawl's id. While the crawl runs, the result has a next link: send the cursor value from that link as cursor to get the pages finished since. fields keeps only the keys you name on each row. Most calls cost 1 credit. Search costs 2 credits per 10 results, a crawl 1 credit a page, and extract 5 credits for most pages. Each result shows its cost in creditsUsed. See pricing.

To give each endpoint its own tool, add toolsets to the URL with a source name, such as https://api.stophy.dev/mcp?toolsets=web. Your agent then also gets scrape, crawlStatus, batchScrape and every other web endpoint as tools.

Network and credentials

The hosted server lives at https://mcp.stophy.dev/mcp (sign in) and https://api.stophy.dev/mcp (no key or an API key). Your client sends requests only there.
Credentials: browser sign-in handled by your MCP client, or an optional STOPHY_API_KEY that you set yourself. The local package also reads an optional STOPHY_MCP_URL.
The plugin manifests (.mcp.json, .grok-plugin/plugin.json, mcp.json, plugin.json) only point at the hosted server. They run no code on your machine.
No telemetry.

License

MIT

来源：mcp/README.md，提交 73d998d

工具

stophy_search_endpoints

Find Stophy endpoints

Start here. Send plain words for the data you want. Returns endpoints with their inputs, filters and cost, ready for stophy_call. Costs nothing.

输入参数querycategory

stophy_describe_endpoint

Describe a Stophy endpoint

Optional. Returns an endpoint's full inputs, filter notes, an example and output fields. Costs nothing.

输入参数id必填

stophy_call

Call a Stophy endpoint

Send an endpoint id and its input. Returns the data and what the call cost.

输入参数id必填inputfieldsmaxTextLength

stophy_scrape

Web Scrape

Scrape any public web page as markdown, HTML, links, a screenshot or JSON. 1 credit for any page, however it is reached. checkPromptInjection adds 4 credits when the scan runs, json with a schema or a prompt, summary, highlights, question, product, menu and changeTracking in json mode add 4 credits each when they return something. CSS rules and attributes add nothing. A PDF adds 1 credit for each page read. Browser actions add nothing. An answer from a stored copy within maxAge costs 1 credit like any page. A page that comes back with a 404 or 410 status is charged like any page. A page that cannot be read costs nothing. creditsUsed shows what the call cost. To get a site's own data as JSON, send the site's page URL with formats [{"type":"json","parser":"<name>"}] and put extra inputs such as page or filters in options. Some sites can only be read this way. Parsers: google-maps, google-maps-place, google-maps-reviews, google-trends, google-scholar, google-patents, google-ads, google-flights, google-hotels, google-play-app, google-play-search, google-play-reviews, upwork-search, upwork-job, craigslist-search, craigslist-listing, careers-jobs, careers-job, indeed-search, indeed-job, realtor-search, realtor-property, zillow-search, zillow-property, expedia-hotels, booking-search, booking-hotel, booking-reviews, airbnb-search, airbnb-listing, airbnb-reviews, youtube-search, youtube-video, youtube-transcript, youtube-comments, youtube-channel, youtube-post, youtube-playlist, reddit-search, reddit-post, reddit-subreddit, reddit-user, reddit-subreddits, instagram-post, instagram-profile, instagram-comments, instagram-transcript, instagram-feed, x-profile, x-post, threads-profile, threads-post, pinterest-search, pinterest-pin, pinterest-board, pinterest-profile, facebook-page, facebook-post, facebook-comments, facebook-marketplace, tiktok-video, tiktok-comments, tiktok-transcript, tiktok-profile, tiktok-hashtag, tiktok-sound, tiktok-search, tiktok-ads, tiktok-shop, tiktok-f

输入参数url必填urlsproxymaxAgeminAgemobileactionsformatsheadersparsersprofiletimeout另有 24 个

stophy_search

Search the web, news, images, scholarly papers, companies and people in one call. 2 credits per 10 results from each ...

输入参数tbssafelimitquery必填maxAgecountrysourcestimeoutlocationcategoriesenterprisehighlights另有 7 个

stophy_map

Web Map

Scrape the list of pages on a website. 1 credit per call. Input: url, search, sitemap, includeSubdomains, ignoreQuery...

输入参数url必填limitsearchsitemaptimeoutlocationalexandriadomainToolsignoreCacheauditMetadataignoreRobotsTxtrobotsUserAgent另有 3 个

stophy_crawl

Web Crawl Start

Crawl a website in the background. 1 credit a page, however it is reached. checkPromptInjection adds 4 credits when t...

输入参数url必填delaylimitpromptsitemapwebhookalexandriadomainToolsexcludePathsincludePathsauditMetadatascrapeOptions另有 12 个

stophy_extract

Web Extract

Pull the fields you name from any public page as JSON, each checked against the page. 5 credits for most pages: the p...

输入参数urlspromptschemaalexandriadomainToolsshowSourcesauditMetadataignoreSitemapscrapeOptionsenableWebSearchignoreRobotsTxtrobotsUserAgent另有 4 个

版本历史

v1.3.1最新Nov 7, 2026
v1.3.0Nov 3, 2026
v2.0.9Oct 10, 2026
v2.0.8Oct 2, 2026
v2.0.6Oct 2, 2026
v2.0.5Oct 2, 2026
v2.0.4Oct 2, 2026
v2.0.3Oct 1, 2026
v2.0.2Sep 30, 2026
v2.0.1Sep 29, 2026
v2.0.0Sep 29, 2026

评分

已满足适用检查项的 90%

该评分衡量此收录条目的完整程度，依据注册中心条目、端点与公开来源计算得出。它不是对服务器本身的评价，也不表示服务器安全或当前可用：可连通性是带日期的检查结果，评分也不包含安装量或用户评价。

近 30 天内可连通必需25 分
已声明认证方式必需10 分
已列出工具必需15 分
描述必需15 分
图标10 分
文档15 分
许可证10 分
请在仓库或注册中心条目中声明许可证。

Stophy

概览

安装

在 SourceWeft 中

交给智能体安装

其他 MCP 客户端

README

Stophy MCP

Connect to the hosted server

Claude

Claude Code

Codex

Cursor

Run it as a local server

Claude Desktop

Tools

Network and credentials

More

License

工具

stophy_search_endpoints

stophy_describe_endpoint

stophy_call

stophy_scrape

stophy_search

stophy_map

stophy_crawl

stophy_extract

stophy_signup

版本历史

评分