Web Content & URL Tools

ai.avaloncompanyv1.0.0更新于 Oct 3, 2026

26 tools: web pages to Markdown, metadata, tech stack, SSL, SEO, domains, email auth and SSO checks.

已验证Streamable HTTP可网页运行Developer ToolsSecurity & MonitoringWeb Search & Scraping

概览

AI 生成的概览

一个托管型 MCP 服务器,提供 26 个工具,可将网页读取为 Markdown、提取元数据与结构,并检查 URL、域名、SEO、邮件认证与 SSO。

功能
它提供 26 个工具,用于抓取并分析公开网页和 URL。页面类工具可将页面转换为干净的 Markdown、提取元数据、抽取标题层级、链接、图片与 JSON-LD、识别技术栈、追踪重定向、检查 TLS 证书与安全响应头、解析 robots.txt、站点地图与订阅源、查询 WHOIS/DNS、审计页面 SEO,并检查最多 100 个链接。它还能评估邮件认证(SPF、DMARC、DKIM、MTA-STS、BIMI)、检查 SAML 与 OpenID Connect 配置,并提供 cron、正则、颜色、Unicode、时区转换等开发者工具,以及 Hyperliquid 永续合约行情数据。
适用场景
当助手需要在不运行浏览器的情况下读取或审计公开网页内容时使用:总结页面、检查站点 SEO、安全响应头、证书、重定向或失效链接、验证邮件或 SSO 配置,或做小型开发者转换。
运行要求
托管在 Apify 上的远程 Streamable HTTP 端点,本地无需运行任何程序。需要 Apify API 令牌,以 Authorization 头按 Bearer 形式发送,或通过 Apify 的 OAuth MCP 网关登录 Apify 账号。需要能访问该端点及目标站点。
安装前请注意
Authorization 头携带 Apify API 令牌,应视为机密,避免分享客户端配置。调用按成功次数通过 Apify 账号计费(检查与实用工具每次 $0.001,页面工具每次 $0.003),失败调用免费。工具只抓取公开的 http/https 地址,拒绝私有或内部地址,部分站点会阻止自动访问。返回的行情数据不构成投资建议。

安装

在 SourceWeft 中

  1. 打开 控制台中的 Web Content & URL Tools,将其添加到工作区。
  2. 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。

其他 MCP 客户端

把它添加到你客户端的 mcpServers 配置中。

{
  "mcpServers": {
    "web-content-url-tools": {
      "type": "http",
      "url": "https://avalonai--web-content-url-tools-mcp.apify.actor/mcp"
    }
  }
}

README

Web Content & URL Tools MCP Server

A remote Model Context Protocol server that gives AI agents 26 tools for reading and checking the web — no browser, no setup, pay only for successful calls.

Read any page as clean Markdown, pull its metadata or structure, detect its tech stack, and check redirects, TLS certificates, security headers, robots.txt, sitemaps, feeds, domains (WHOIS/DNS), SEO, broken links, email authentication (SPF, DMARC, DKIM) and single sign-on (SAML, OpenID Connect). It also includes everyday developer utilities (cron, regex, color, Unicode, time zones) and Hyperliquid perpetuals market data.

Connect

Option 1 — Standby URL + Apify API token

Use your Apify API token as a Bearer token. Works in Claude Desktop, Claude Code, Cursor, VS Code, Windsurf and any MCP client that supports remote servers with headers.

json
{  "mcpServers": {    "web-content-url-tools": {      "url": "https://avalonai--web-content-url-tools-mcp.apify.actor/mcp",      "headers": { "Authorization": "Bearer <YOUR_APIFY_TOKEN>" }    }  }}

Option 2 — OAuth via mcp.apify.com (no token to copy)

Clients that support MCP OAuth can connect through Apify's hosted MCP gateway and sign in with their Apify account in the browser:

https://mcp.apify.com?tools=avalonai/web-content-url-tools-mcp

Claude Code example:

bash
claude mcp add --transport http web-content-url-tools "https://mcp.apify.com?tools=avalonai/web-content-url-tools-mcp"

Tools

Web pages and URLs

ToolWhat it returnsPrice per call
to_markdownClean, LLM-ready Markdown of the main content (menus, ads and footers removed), word/character counts, language$0.003
get_metadataTitle, description, preview image, site name, canonical URL, favicon, author, publish time, Open Graph / Twitter tags$0.003
extract_structureHeading outline, internal/external links, images, tables and JSON-LD$0.003
detect_technologiesCMS, framework, e-commerce platform, CDN, analytics and payment tools, with evidence$0.003
trace_redirectsEvery redirect hop (HTTP and meta refresh) to the final URL — expands short links$0.001
check_sslCertificate issuer, expiry, days remaining, SANs, and whether browsers trust it$0.001
grade_security_headersA+ to F grade for security headers, information leaks and insecure cookies$0.001
parse_robotsrobots.txt rules per user agent, sitemaps, and whether a path may be crawled$0.001
parse_sitemapURLs from a sitemap or sitemap index (auto-discovered, gzip supported)$0.001
read_feedRSS, Atom or RDF feed as JSON (auto-discovered from a website URL)$0.001
lookup_domainWHOIS (RDAP) registrar, dates, domain age, nameservers, DNS records, MX/SPF/DMARC$0.001
audit_seoOn-page SEO score and grade with prioritised issues and fixes$0.003
check_broken_linksUp to 100 links checked: broken, redirected and bot-blocked reported separately$0.003

Email and identity diagnostics

ToolWhat it returnsPrice per call
check_email_domainMX, SPF (10-lookup limit), DMARC, DKIM, MTA-STS, TLS-RPT and BIMI graded pass/warn/fail with fixes and a 0-100 score$0.003
analyze_saml_metadataSAML metadata (XML or URL): entity ID, endpoints and bindings, NameID formats, signing certificate expiry and issues$0.001
check_oidc_providerOpenID Connect discovery and JWKS: issuer match, HTTPS, PKCE S256, ID token algorithms, key sizes, certificate expiry$0.001

Developer utilities

ToolWhat it returnsPrice per call
parse_cronValidity, plain-English explanation, previous and next run times in any time zone$0.001
test_regexEvery match with positions and groups, plus find-and-replace; runaway patterns stop in 0.1 s$0.001
convert_colorHEX, RGB, HSL, HSV, CMYK, best text color, contrast and palettes$0.001
check_color_contrastWCAG 2 contrast ratio with AA/AAA pass/fail$0.001
inspect_unicodePer-character code point, name and bytes; grapheme count; normalization forms$0.001
convert_timezoneOne time in many IANA zones with offsets, abbreviations and DST$0.001
list_timezonesIANA time zone names, optionally by region$0.001

Hyperliquid perpetuals market data

ToolWhat it returnsPrice per call
hyperliquid_marketsMark/oracle price, 24h change, hourly and annualized funding, open interest, volume$0.001
hyperliquid_funding_historyHourly funding history (up to 30 days) with average and annualized rate$0.001
hyperliquid_candlesOHLCV candles from 1 minute to 1 week$0.001

Hyperliquid data comes from Hyperliquid's public market data. This server is not affiliated with Hyperliquid, and nothing it returns is financial advice.

Pricing

Pay per successful tool call through your Apify account: $0.001 for check and utility tools, $0.003 for page tools. Failed calls (invalid URL, blocked site, timeout) are free and return a clear error reason.

Good to know

  • Pages are read from the HTML the server sends; sites that render everything in the browser return little text and say so in a warning field.
  • Some sites block automated access; those calls return target_blocked and are not charged.
  • Only public http/https addresses are fetched; private and internal addresses are refused.

This repository

This is the source of the Apify Actor that hosts the MCP server. It is a thin connector: each tool forwards the request to the Web Content API and returns the JSON result.

.actor/        Apify Actor definition (Standby mode, MCP path /mcp, pay-per-event pricing)src/main.py    MCP server (FastMCP, Streamable HTTP) and tool definitionsDockerfile     Container image

You do not need to run it yourself — connect to the hosted endpoint above.

License

MIT © Avalon Company

来源:README.md,提交 44579ed

工具

0
工具元数据尚未被收录。

版本历史

1
  1. v1.0.0最新Oct 3, 2026