PDF to Text & Markdown: Bulk Extractor for AI and RAG

Paste PDF links and get clean, readable text, ready for ChatGPT, Claude, a RAG index or plain analysis.

• Reading order rebuilt: words joined into lines, lines into paragraphs, wrapped lines merged
• Markdown with headings: larger-font lines become # / ## headings
• Metadata: title, author, subject, creator app, created and modified dates
• Counts: total pages, pages with text, words and approximate tokens
• Per-page text (optional), handy for citations ("see page 4")
• Scanned PDFs flagged: pages that are only images are detected and not charged
• Priced per page with text, no start fee

Who it's for

• AI and RAG builders: index reports, papers, manuals and contracts.
• Researchers and analysts: pull text from many PDFs into one table.
• Operations teams: read the text of forms and documents in bulk.

Input

Option	Default	What it does
PDF links	–	Direct links to PDF files
Output format	Markdown	Markdown, plain text or both
Also give text per page	off	Adds a pages array
Max pages per PDF	200	Read at most this many pages

Output

Real result (first 3 pages of a well-known paper, shortened):

{
  "url": "https://arxiv.org/pdf/1706.03762",
  "status": "ok",
  "title": null,
  "createdAt": "2024-04-10",
  "totalPages": 15,
  "pagesExtracted": 3,
  "pagesWithText": 3,
  "likelyScanned": false,
  "wordCount": 1315,
  "markdown": "…\n\n# Attention Is All You Need\n\n…\n\nAbstract\n\nThe dominant sequence transduction models are based on …"
}

status	meaning
ok	Extracted.
unreachable	The link didn't load. Not charged.
not_pdf	The link isn't a PDF. Not charged.
password_protected / unreadable	Can't be opened. Not charged.
too_large	Over 100 MB. Not charged.

Use it from your AI agent (MCP)

Claude, Cursor and other MCP clients can call this tool directly through Apify's MCP server. Add it as an MCP server / connector:

{
  "mcpServers": {
    "apify": {
      "url": "https://mcp.apify.com?tools=swiftkit/pdf-text",
      "headers": { "Authorization": "Bearer <YOUR_APIFY_TOKEN>" }
    }
  }
}

Then just ask, for example: "Extract the text of this PDF and list its main findings: https://…/report.pdf". Each call is billed like a normal run.

Pricing

You pay per page that has text. Scanned (image-only) pages and failed PDFs are free. See the Pricing tab.

Limits, honestly

• No OCR. Scanned PDFs (photos of paper) have no text layer; they're flagged with likelyScanned: true and not charged, but no text comes out.
• Tables are flattened into lines of text, and complex layouts like multi-column forms can mix columns. Ordinary single-column documents, reports and most papers come out clean.
• Title and author come from the PDF's own metadata, which many PDFs leave empty.
• Links must point straight at the PDF file (not a viewer page or a login).

More tools from SwiftKit

• Website to Markdown for AI: clean page content as Markdown, with token counts and RAG chunks
• Sitemap Extractor & Broken Link Checker: every URL from a site’s sitemaps, plus 404 and redirect checks
• RSS Feed Reader: news, blogs and podcasts from any feed or website, only new items

Questions?

Open an issue on the Issues tab.