{
  "listings": [
    {
      "slug": "apify-mcp",
      "name": "Apify MCP Server",
      "company": "Apify Technologies s.r.o.",
      "grade": "A",
      "fit": "yes",
      "why": "The default apify--web-fetch tool returns Markdown narrowed to the main article or with boilerplate stripped, renders JavaScript, reads PDFs, costs $1 per 1,000 fetches with failures and retries unbilled, and caps pages at 10 MB and two minutes.",
      "source": "https://apify.com/apify/web-fetch",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/apify-mcp"
    },
    {
      "slug": "context-dev",
      "name": "Context.dev",
      "company": "Context Dev Inc.",
      "grade": "A",
      "fit": "yes",
      "why": "POST /v1/web/scrape with formats.markdown and sharedParams.mainContentOnly returns Markdown for 1 credit, handles PDFs and 40 file types, and every output carries success and error_code (WEBSITE_BLOCKED for a challenge) even on HTTP 200. No token cap.",
      "source": "https://docs.context.dev/api-reference/web-scraping/scrape.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/context-dev"
    },
    {
      "slug": "crw",
      "name": "fastCRW",
      "company": "Nanofy OÜ",
      "grade": "B",
      "fit": "yes",
      "why": "POST /v1/scrape defaults to markdown with onlyMainContent true, renderJs auto-detects SPAs at 1 credit (chrome 2), maxLength truncates at about 15,000 characters with startIndex to continue, parsers handle PDF, and anti-bot walls come back flagged rather than as page text.",
      "source": "https://docs.fastcrw.com/llms-full.txt",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/crw"
    },
    {
      "slug": "microlink",
      "name": "Microlink",
      "company": "Microlink Labs, S.L.",
      "grade": "B",
      "fit": "yes",
      "why": "data.markdown.attr=markdown returns the page as Markdown in one GET, prerender auto renders JavaScript when needed, and a selector scopes noisy pages; the MCP's microlink_markdown wraps it. 25 requests a day free, no token cap, no main-content default.",
      "source": "https://microlink.io/docs/guides/content-conversion/url-to-markdown.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/microlink"
    },
    {
      "slug": "olostep",
      "name": "Olostep",
      "company": "Olostep Technologies Inc.",
      "grade": "B",
      "fit": "yes",
      "why": "/v1/scrapes with formats markdown returns markdown_content, remove_css_selectors defaults to stripping nav, footer and script noise, every request is JS-rendered on residential IPs at 1 credit, PDFs are handled, and credits_consumed is in the response. No token cap.",
      "source": "https://docs.olostep.com/features/scrapes.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/olostep"
    },
    {
      "slug": "outcrawl",
      "name": "Outcrawl",
      "company": "Aeonmind LLC (trading as Outcrawl)",
      "grade": "B",
      "fit": "yes",
      "why": "POST /v1/scrape with formats markdown returns the main content by default (onlyMainContent false for the whole page) at 1 credit, with usage in every response and warning codes such as blocked_interstitial and empty_capture. No token cap; PDFs not mentioned.",
      "source": "https://outcrawl.ai/docs/scrape/overview",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/outcrawl"
    },
    {
      "slug": "oxylabs-web-scraper",
      "name": "Oxylabs Web Scraper API",
      "company": "oxylabs, UAB",
      "grade": "B",
      "fit": "yes",
      "why": "Web API POST /v1/scrape with output markdown returns Markdown, run_js is optional and cheaper off, and the endpoint handles rendering and access challenges itself; set a 150-second client timeout. No main-content option or token cap, and a 4xx page is billed.",
      "source": "https://developers.oxylabs.io/products/web-api/scrape.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/oxylabs-web-scraper"
    },
    {
      "slug": "scrapegraphai",
      "name": "ScrapeGraphAI",
      "company": "ScrapeGraphAI, Inc.",
      "grade": "B",
      "fit": "yes",
      "why": "scrape with formats markdown costs 1 credit, mode reader or prune strips chrome, fetchConfig mode auto renders JavaScript when needed, PDFs convert at 1 credit a page (25-page default cap). Response carries status success or an error.",
      "source": "https://docs.scrapegraphai.com/services/scrape.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/scrapegraphai"
    },
    {
      "slug": "scrapfly",
      "name": "Scrapfly",
      "company": "Joam Intelligence, LLC",
      "grade": "BB",
      "fit": "yes",
      "why": "format=markdown with the only_content option extracts the main content, no_links optional, 1 credit on datacenter and +5 for render_js. The X-Scrapfly-Api-Cost header reports the charge and ERR codes mark failures.",
      "source": "https://scrapfly.io/docs/scrape-api/getting-started",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/scrapfly"
    },
    {
      "slug": "scrapingbee",
      "name": "ScrapingBee",
      "company": "VostokInc SAS",
      "grade": "B",
      "fit": "yes",
      "why": "return_page_markdown=true returns the main page content as Markdown; render_js=false cuts a call from 5 credits to 1, Spb-cost reports the charge and non-2xx targets other than 404 come back as a 500, not a 200 body.",
      "source": "https://www.scrapingbee.com/llms/html-api.txt",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/scrapingbee"
    },
    {
      "slug": "spider-cloud",
      "name": "Spider",
      "company": "BAGELMEN LLC",
      "grade": "BB",
      "fit": "yes",
      "why": "/scrape with return_format markdown and readability true strips boilerplate; request smart renders only when needed; costs come back per page and a 200 API response carries the page's own status. Works keyless at a low rate.",
      "source": "https://spider.cloud/llms-full.txt",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/spider-cloud"
    },
    {
      "slug": "webscraping-ai",
      "name": "WebScraping.AI",
      "company": "Urlooker LLC",
      "grade": "B",
      "fit": "yes",
      "why": "/text returns clean Markdown with boilerplate stripped and headings, lists and links kept; js=false drops the cost from 5 credits to 1, requests cap at 30 s, and a block or challenge is a 500 with error_code, never a 200 body.",
      "source": "https://webscraping.ai/docs.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/webscraping-ai"
    },
    {
      "slug": "zyte",
      "name": "Zyte API",
      "company": "Zyte Group Limited",
      "grade": "BB",
      "fit": "yes",
      "why": "markdown: true returns the page as Markdown scoped to the main content by default, from httpResponseBody or browserHtml when rendering is needed; a 404 is reported in statusCode and bans come back as 520, not a 200 body.",
      "source": "https://docs.zyte.com/zyte-api/usage/reference.md",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/zyte"
    },
    {
      "slug": "bright-data",
      "name": "Bright Data",
      "company": "Bright Data Ltd.",
      "grade": "C",
      "fit": "caveat",
      "why": "scrape_as_markdown uses the Fetch API and returns Markdown with headings, lists and links for 1 credit, but it does no JavaScript rendering, times out at 45 seconds and has no main-content option; JavaScript pages need the browser tools.",
      "source": "https://docs.brightdata.com/products/mcp-server/tools.md",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/bright-data"
    },
    {
      "slug": "crawl4ai",
      "name": "Crawl4AI",
      "company": "CONTEXT4AI PTE LTD",
      "grade": "D",
      "fit": "caveat",
      "why": "POST /scrape returns Markdown, picks cache, HTTP or browser itself, and a blocked or login-walled page is a 403 that isn't charged. There is no main-content or token option, PDFs aren't mentioned, and acc-03 says Crawl4AI missed shadow DOM content.",
      "source": "https://crawl4ai.com/llms.txt",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/crawl4ai"
    },
    {
      "slug": "decodo",
      "name": "Decodo Web Data API",
      "company": "UAB \"Data troops\"",
      "grade": "B",
      "fit": "caveat",
      "why": "markdown: true on /scrape converts the page and headless: html renders JavaScript, but there is no main-content filter, requests default to the premium pool at twice the price, and a 403 or 404 page with content is billed.",
      "source": "https://help.decodo.com/docs/web-data-api-scrape-parameters.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/decodo"
    },
    {
      "slug": "diffbot",
      "name": "Diffbot",
      "company": "Diffbot Technologies Corp.",
      "grade": "C",
      "fit": "caveat",
      "why": "Article API returns the clean article as text and html fields in JSON after Chromium rendering, 1 credit a page, with a 30-second fetch timeout. No Markdown format, the text field drops links, and a 429 is the only documented failure read.",
      "source": "https://www.diffbot.com/docs/extract/article",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/diffbot"
    },
    {
      "slug": "firecrawl-mcp",
      "name": "Firecrawl MCP",
      "company": "SideGuide Technologies, Inc. (d/b/a Firecrawl)",
      "grade": "A",
      "fit": "caveat",
      "why": "Scrape returns Markdown with links, onlyMainContent defaults true, PDFs parse in the same call and success plus metadata.statusCode flag bad pages. acc-32 shows onlyMainContent dropping the body on CivicPlus pages and acc-03 shadow DOM content missed; no token cap.",
      "source": "https://docs.firecrawl.dev/features/scrape.md",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/firecrawl-mcp"
    },
    {
      "slug": "jevcrawl",
      "name": "jevcrawl",
      "company": "Rayen Mlayeh",
      "grade": "B",
      "fit": "caveat",
      "why": "scrape(url) without a goal returns the whole page's content with navigation and footers removed and makes no paid Jev call, but it reads plain HTML unless the js extra is installed locally, and the output format isn't described as Markdown.",
      "source": "https://github.com/RayenMlayeh/jevcrawl/blob/main/README.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/jevcrawl"
    },
    {
      "slug": "jina-reader",
      "name": "Jina Reader",
      "company": "Jina AI GmbH",
      "grade": "C",
      "fit": "caveat",
      "why": "r.jina.ai returns Markdown with links, readability main content, x-max-tokens truncation, native PDF and x-with-shadow-dom. acc-11 cut a Wikipedia page to 3 to 5k tokens. acc-30 got HTTP 200 with a Cloudflare page body and acc-14 still needed a noise filter.",
      "source": "https://github.com/jina-ai/reader/blob/main/README.md",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/jina-reader"
    },
    {
      "slug": "scrape-do",
      "name": "Scrape.do",
      "company": "PACKEND, LLC",
      "grade": "E",
      "fit": "caveat",
      "why": "output=markdown converts the whole text/html page; no main-content option and the docs don't say whether links survive. Datacenter 1 credit, render=true 5, 4 MB body cap, timeout 5 to 120 s.",
      "source": "https://scrape.do/documentation/api-response/response-output.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/scrape-do"
    },
    {
      "slug": "scrapeless",
      "name": "Scrapeless",
      "company": "NST LABS TECH LTD.",
      "grade": "C",
      "fit": "caveat",
      "why": "Crawler scrape takes formats markdown and onlyMainContent, but it is a job polled through a status endpoint, and billing is proxy traffic at $1.8 a GB plus $0.09 a browser hour rather than a known per-page price.",
      "source": "https://apidocs.scrapeless.com/api-17509002.md",
      "confidence": "low",
      "url": "https://www.anchorterminal.com/tools/scrapeless"
    },
    {
      "slug": "scrapeops",
      "name": "ScrapeOps Proxy API Aggregator",
      "company": "ScrapeOps Limited",
      "grade": "D",
      "fit": "caveat",
      "why": "The aggregator returns raw HTML; Markdown comes only from llm_extract=true with llm_extract_response_type=markdown at 25 extra credits. render_js is 10 credits (render_js_cheap 5).",
      "source": "https://scrapeops.io/docs/web-scraping-proxy-api-aggregator/advanced-functionality/llm-extract/",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/scrapeops"
    },
    {
      "slug": "scraperapi",
      "name": "ScraperAPI",
      "company": "ScraperAPI LLC",
      "grade": "D",
      "fit": "caveat",
      "why": "output_format=markdown keeps links, but the docs' own Amazon sample starts with the department menu and sign-in links: the whole page converts, no main-content option. 1 credit plain, 10 rendered, 70 s retry window.",
      "source": "https://docs.scraperapi.com/responses-and-formats/output-formats/llm-output-formats.md",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/scraperapi"
    },
    {
      "slug": "scrapewise",
      "name": "Scrapewise",
      "company": "BEBOTECH OÜ",
      "grade": "B",
      "fit": "caveat",
      "why": "fetch_page returns the text or HTML of one page for usually 5 credits (at most 25, so render or residential tiers), with quote_only to see the price first. No Markdown format or main-content statement.",
      "source": "https://scrapewise.ai/llms.txt",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/scrapewise"
    },
    {
      "slug": "scrapingant",
      "name": "ScrapingAnt",
      "company": "DATAANT SP. Z O.O.",
      "grade": "E",
      "fit": "caveat",
      "why": "/v2/markdown converts the whole page with html2text, links kept inline; the docs say navigation, footers and cookie banners are not removed, and a target 404 still returns HTTP 200. Browser default costs 10 credits, browser=false 1.",
      "source": "https://docs.scrapingant.com/llm-markdown",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/scrapingant"
    },
    {
      "slug": "scrapingdog",
      "name": "Scrapingdog",
      "company": "OpenData Labs Private Limited",
      "grade": "E",
      "fit": "caveat",
      "why": "/scrape takes markdown=true for clean text (skill) and dynamic=false drops rendering from 5 credits to 1; no main-content statement, no sample, and a 60-second timeout.",
      "source": "https://www.scrapingdog.com/agent-onboarding/SKILL.md",
      "confidence": "low",
      "url": "https://www.anchorterminal.com/tools/scrapingdog"
    },
    {
      "slug": "zenrows",
      "name": "ZenRows",
      "company": "ZENROWS, S.L.",
      "grade": "A",
      "fit": "caveat",
      "why": "response_type=markdown keeps headings, links and tables, but the docs say navigation, footers and sidebars stay in, and a 200 is charged whatever the Markdown holds. 1 credit plain, 5 with js_render.",
      "source": "https://docs.zenrows.com/fetch/features/markdown.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/zenrows"
    },
    {
      "slug": "browseract",
      "name": "BrowserAct",
      "company": "Hongkong Ziniao Technology Co., Limited",
      "grade": "D",
      "fit": "no",
      "why": "No URL-to-Markdown call. A bot must be built first (300 to 1,500 credits, with a confirmation step inside 60 minutes) and then run at 10 to 30 credits; output arrives as result files after the task finishes.",
      "source": "https://docs.browseract.com/bot/agent-built/agent-built.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/browseract"
    },
    {
      "slug": "qovia",
      "name": "Qovia",
      "company": "Qovia",
      "grade": "D",
      "fit": "no",
      "why": "No URL fetch; every endpoint takes a platform identifier and returns that platform's JSON.",
      "source": "https://qovia.ai/llms-full.txt",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/qovia"
    },
    {
      "slug": "scrapeatlas",
      "name": "ScrapeAtlas",
      "company": "OD. Media \u0026 Technology LLC",
      "grade": "B",
      "fit": "no",
      "why": "No URL fetch endpoint. Every operation is a platform-specific JSON endpoint; an arbitrary page cannot be read as text.",
      "source": "https://scrapeatlas.com/docs/llms.txt",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/scrapeatlas"
    },
    {
      "slug": "soldfetch",
      "name": "SoldFetch",
      "company": "SoldFetch",
      "grade": "E",
      "fit": "no",
      "why": "No page fetch; the API returns eBay listing records only.",
      "source": "https://soldfetch.com/docs/llms-full.txt",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/soldfetch"
    }
  ],
  "marks": "yes: the docs show it does the job as the success test says. caveat: it does the job with a limit an agent should know. no: it doesn't do this job. unchecked: we couldn't read it.",
  "meta": {
    "attribution": "Anchor Terminal (https://www.anchorterminal.com)",
    "attributionRequired": true,
    "credit": "CC BY 4.0, credit Anchor Terminal and link the page or API address you took it from (https://www.anchorterminal.com/terms/#data)",
    "docs": "https://www.anchorterminal.com/docs/",
    "generatedAt": "2026-11-16",
    "license": "CC-BY-4.0",
    "method": "https://www.anchorterminal.com/benchmark/",
    "methodology": "0.8",
    "openapi": "https://www.anchorterminal.com/openapi.json",
    "preview": false,
    "run": "2026-10-01",
    "runLabel": "October 2026 research run",
    "terms": "https://www.anchorterminal.com/terms/#data"
  },
  "useCase": {
    "accounts": 14,
    "after": [
      "document-extraction.table-from-text",
      "llm.summarise"
    ],
    "area": "web-data",
    "asks": [
      "get me the text of this page",
      "read this URL",
      "what does this page say",
      "fetch this article as markdown"
    ],
    "before": [
      "search.find-url",
      "chat.user-pastes-url"
    ],
    "blockers": [
      "bot wall answered as HTTP 200 with a challenge page body",
      "main-content filter strips the content the user wanted",
      "page is JavaScript built inside shadow DOM"
    ],
    "capabilities": [
      "web.scrape",
      "web.fetch",
      "scraping.js-render"
    ],
    "category": "web-scrapers",
    "categoryName": "Web scraping \u0026 crawling",
    "checked": "2026-11-12",
    "companies": {
      "fit": 13,
      "fitWithLimit": 15,
      "tagged": 55
    },
    "competition": "crowded",
    "fits": [
      "apify-mcp",
      "context-dev",
      "crw",
      "microlink",
      "olostep",
      "outcrawl",
      "oxylabs-web-scraper",
      "scrapegraphai",
      "scrapfly",
      "scrapingbee",
      "spider-cloud",
      "webscraping-ai",
      "zyte"
    ],
    "fitsWithLimit": [
      "bright-data",
      "crawl4ai",
      "decodo",
      "diffbot",
      "firecrawl-mcp",
      "jevcrawl",
      "jina-reader",
      "scrape-do",
      "scrapeless",
      "scrapeops",
      "scraperapi",
      "scrapewise",
      "scrapingant",
      "scrapingdog",
      "zenrows"
    ],
    "id": "web-scrapers.page-to-markdown",
    "listings": {
      "caveat": 15,
      "no": 4,
      "notRead": 1,
      "tagged": 56,
      "unchecked": 0,
      "yes": 13
    },
    "name": "Read one page as text an agent can use",
    "status": "draft",
    "success": "Markdown or clean text of the main content with links kept, in one call. Measured: under 10,000 tokens for a typical article, within 15 seconds, and the body is not a challenge page.",
    "url": "https://www.anchorterminal.com/api/v1/use-cases/web-scrapers.page-to-markdown.json",
    "weight": 26,
    "weightBasis": "14 of 35 accounts; the job most often described on Hacker News (Jina Reader and Firecrawl threads) and in the PageMap benchmark"
  }
}
