Web Scraper
Scrape, crawl & extract web data
On it. Running Web Scraper now and I'll bring the results straight back here.
Try asking
About
Web Scraper turns any website into structured, readable data. It handles JavaScript-heavy pages, anti-bot protection, geographic targeting, and AI-powered data extraction — all without writing any code. Whether you need the text from a single page, a full site crawl, or a structured dataset extracted to a schema, it's one tool.
Standard web scrapers break on modern sites that require JavaScript to render. This one renders fully, bypasses bot detection, supports custom request headers, and can extract specific data fields using AI so you get exactly what you asked for, not raw HTML. Checked 2026-08-26 against this manifest.
What you can do
- scrape_page — scrape a single page with JS rendering; returns markdown, HTML, filtered content, or JSON responses
- crawl_site — recursively crawl a site up to a depth and page limit, collecting all content
- map_site — fast URL discovery without content, useful before deciding what to crawl
- extract_data — AI-powered structured extraction — describe what you want or pass a JSON schema
- search_web — search the web and optionally scrape the top results in one call
- stealth_scrape / stealth_crawl — enhanced anti-bot bypass for heavily protected pages
Who it's for
Researchers aggregating content from multiple sites. Data teams building datasets from web sources. Developers integrating web content into pipelines. Marketers monitoring competitor offers and messaging. Analysts tracking content changes over time.
How to use it
- Start with scrape_page for a single URL — add onlyMainContent to strip navigation and footers
- Use map_site to discover URLs on a domain before deciding what to crawl
- Use crawl_site for recursive multi-page collection with depth and path filters
- Use extract_data when you want specific fields — pass a schema or describe what to extract in plain English
- If a page blocks the ordinary request, use stealth_scrape for enhanced routing
- For long React/SPA pages or competitor research, set fullContent to true or increase maxPreviewChars so the model-facing response includes more than the compact preview
fetch_page adds browser controls, cookies and selectors.
Getting started
Try scrape_page with "https://example.com" and a bounded preview. Representative manifest-schema shape: success, data [redacted]. Complete rendering is not guaranteed; timing varies.
Render a public page and return Markdown, HTML or text, with browser…
Run one approved marketplace-product, marketplace-review, or…
Scrape a single web page with full JavaScript rendering, anti-bot…
Recursively crawl a website starting from a URL, following links up to…
Render a public page and return Markdown, HTML or text, with browser, selector and cookie controls.
Run one approved marketplace-product, marketplace-review, or public-profile collection through an explicitly selected configured provider; arbitrary dataset and actor IDs are blocked.
Scrape a single web page with full JavaScript rendering, anti-bot bypass, and configurable output formats. Supports markdown, HTML, and content filtering by CSS tags.
Recursively crawl a website starting from a URL, following links up to a configurable depth and page limit. Returns scraped content for all discovered pages.
Quickly discover all URLs on a website using sitemaps and link analysis. Returns a flat list of URLs without scraping content. Optionally filter by keyword.
Use AI to extract structured data from one or more web pages. Provide a JSON Schema for typed output or a natural language prompt for flexible extraction.
Search the web using a query and optionally scrape the content of each result page. Returns search results with titles, URLs, and snippets, plus full page content when scraping is enabled.
Scrape a single bot-protected web page using enhanced residential proxies, geo-targeted IPs, and extended rendering wait times. Bypasses Cloudflare, Akamai, DataDome, and similar anti-bot systems.
Recursively crawl a bot-protected website using enhanced proxies on every page. Bypasses anti-bot systems across the entire crawl, with geo-targeted IPs and extended rendering.
- Added an explicit rendered-page source with selectors, cookies and browser controls
- Concise responses now honor fullContent and maxPreviewChars, returning the requested page text without a second call
- Structured extraction now uses synchronous Firecrawl v2 JSON scraping, with ordered results per URL and provider-credit accounting.
- map_site now falls back to the site's own sitemaps when the main mapper is slow or unavailable, returning source "sitemap" at no provider cost
- scrape_page now uses TinyFish by default for faster, lower-cost page fetches; proxy, geo-targeting and mobile options still use Firecrawl
- scrape_page can now use TinyFish as an alternative scraping provider via the provider option
- Added curated Bright Data snapshots and pinned Apify tasks for three approved data jobs; arbitrary dataset and actor execution remains blocked
- Added fullContent and maxPreviewChars controls to scrape_page and stealth_scrape for long React/SPA pages and competitor research
- scrape_page now handles JSON API endpoints — returns parsed JSON and markdown instead of empty content
- Added headers parameter to scrape_page for custom request headers (e.g. Authorization)
- Added stealth_scrape and stealth_crawl skills for bot-protected websites
- Added proxy, country, and languages parameters to scrape_page and crawl_site for anti-bot bypass and geo-targeting
- Added subtitle, expanded description, and agent instructions
- Initial release
Powered by
Related
Guides
- ChatGPT MCP Not Working: Fixes That Match the ProductAdd Toolrouter from the plugin directory. If custom MCP fails, check Developer mode, the chat toggle, and the server URL.
- How to Add MCP on Claude TeamOn Claude Team, only an Owner adds a custom connector. Members then Connect. URL for ToolRouter is https://api.toolrouter.com/mcp/anthropic.
- How to Add MCP to CursorAdd ToolRouter to Cursor with a one-click link or mcp.json. Local npx or a remote HTTPS URL. Tools show in the agent list after restart.
- How to Scrape a Website in ChatGPTUse a read/fetch ToolRouter MCP app to scrape websites in ChatGPT Pro, Business or Enterprise/Edu web with structured output.
Questions
Can it handle JavaScript-heavy sites?
Yes. JavaScript rendering is built in, so dynamic pages are part of the normal workflow.
Can it crawl and discover new URLs?
Yes. It can scrape a page, crawl sites, and discover URLs in one tool.
What formats can I extract?
You can pull markdown, HTML, or typed data depending on how structured you want the result to be.
Does it help with anti-bot pages?
Yes. Anti-bot bypass is part of the web access surface for the tool.