Website to Markdown for LLM & RAG – Fast Site Crawler

Crawl a website or docs section and get clean Markdown per page, with title, language, word and token counts, content-hash dedup and optional overlapping chunks ready for vector databases. No browser, so it is fast and cheap.

Why this actor

• HTTP-only crawler = ~10x cheaper than browser-based RAG crawlers; built-in chunking and dedup; stays under the start path.
• Pay only for results: $1.00 per 1,000 pages, no start fee, failed lookups are free.
• Robust: automatic retries with exponential backoff, respects Retry-After, per-item error rows instead of crashed runs.
• Clean, flat JSON/CSV/Excel output ready for Google Sheets, Make, Zapier, n8n, Clay, or the Apify API.

Use cases

• Build RAG knowledge bases from docs sites
• Feed custom GPTs/assistants with a company's website
• Keep a vector DB fresh on a schedule
• Content audits and migrations

Input

Field	Type	Description
startUrls	array	Site or section to crawl (e.g. https://docs.example.com/).
maxPagesPerSite	integer	Crawl budget per start URL.
maxDepth	integer	How many clicks away from the start URL.
stayUnderStartPath	boolean	Only follow links under the start URL path (e.g. /docs/).
includeRegex	string	Only output pages matching this regex.
excludeRegex	string	Never visit URLs matching this regex (e.g. /tag/
chunkWords	integer	If >0, also output overlapping chunks for vector databases.
chunkOverlapWords	integer	Overlap between chunks.
maxItems	integer	0 = no limit. Caps cost.

Example input:

{
  "startUrls": [
    "https://docs.apify.com/academy/web-scraping-for-beginners"
  ],
  "maxPagesPerSite": 5,
  "maxDepth": 1,
  "chunkWords": 300
}

Output

Each dataset item looks like this (real sample from a test run, long fields trimmed):

{
  "url": "https://docs.apify.com/academy/scraping-basics-javascript?legacy-js-course=/",
  "startUrl": "https://docs.apify.com/academy/web-scraping-for-beginners",
  "depth": 0,
  "title": "Web scraping basics for JavaScript devs | Academy | Apify Documentation",
  "description": "Learn how to use JavaScript to extract information from websites in this practical course, starting from the absolute basics.",
  "language": "en",
  "markdown": "# Web scraping basics for JavaScript devs\n\n**Learn how to use JavaScript to extract information from websites in this practical course, starting from the absolute basics.**\n\nIn this course we'll use JavaScript to create an application for w\u2026",
  "wordCount": 995,
  "approxTokens": 1323,
  "contentHash": "a05d980e413fc565dcec10927a371f95",
  "crawledAt": "2026-10-14T23:42:37+00:00",
  "chunks": [
    "# Web scraping basics for JavaScript devs **Learn how to use JavaScript to extract information from websites in this practical course, starting from the absolute basics.** In this course we'll use JavaScript to create an application for wat\u2026",
    "is web scraping good for, and what career opportunities does it enable for you? ### Why learn scraping The internet is full of useful data, but most of it isn't offered in a structured way that's easy to process programmatically. That's why\u2026",
    "building personal tools, such as your own real estate or rare sneakers watchdog. - Companies can hire you to build custom scrapers mining data important for their business. - Become an invaluable asset to data journalism, data science, or n\u2026",
    "\u2026"
  ]
}

Rows that could not be processed contain an error field and are not charged.

Pricing

Pay-per-event: $0.001 per page (= $1.00 / 1,000). Use Max results to cap spend. Apify's free plan credit covers thousands of results per month.

Integrations & API

Run on a schedule in Apify Console, call via the Apify API / Python & JS clients, or connect to Make, Zapier, n8n, Google Sheets, Slack and webhooks. Works as a tool for AI agents via Apify MCP.

FAQ

Is this an official tool? No. It is an independent actor and is not affiliated with, sponsored or endorsed by the data source or any trademark owner named here.

Is it legal? The actor only reads publicly available data from public endpoints, without logins, and does not bypass paywalls or access controls. You are responsible for using the data in line with applicable laws (e.g. GDPR) and the source sites' terms.

Something broke? Open an issue in the Issues tab; we typically respond within 24 hours and ship fixes fast.

Polite crawling

The crawler reads each site's robots.txt and skips any URL it disallows (honoring Crawl-delay, capped at 10 s). A start URL blocked by robots.txt returns a free error row explaining why.