PDF Text Extractor — Text & Metadata from Any PDF URL

Extract clean plain text and full document metadata from any PDF — just paste the URLs. No API keys, no login, no proxies, no OCR setup. Point it at a list of PDF links and get back structured, LLM-ready text plus title/author/date/page metadata for every file.

Built for developers and data teams who need to turn scattered PDF documents — research papers, financial reports, whitepapers, manuals, invoices, legal filings — into clean structured data at scale.

Why this Actor

• Zero configuration — no credentials, no proxy budget, no browser. Paste URLs, press run.
• RAG-ready output — optional includePageTexts mode returns per-page text so you can chunk documents for embeddings without re-splitting.
• Resilient by design — unreadable or password-protected files are skipped and logged, never crashing the run. A batch of 500 PDFs with 3 bad links still returns 497 clean items.
• Handles big documents — multi-hundred-page files up to 100 MB each, including "secured" PDFs encrypted with an empty password.

How it works

Give it a list of direct PDF URLs. For each file the Actor downloads it, parses it with a battle-tested PDF engine, and outputs one dataset item per PDF with the full extracted text and all document metadata. Set maxPagesPerPdf to cap long files, or includePageTexts: true to also get an array with each page's text separately.

Input

{
  "pdfUrls": [
    "https://arxiv.org/pdf/1706.03762",
    "https://bitcoin.org/bitcoin.pdf"
  ],
  "maxPagesPerPdf": 0,
  "includePageTexts": false
}

Field	Type	Description
pdfUrls	array	Direct URLs of PDF files (required)
maxPagesPerPdf	integer	Max pages to extract per PDF; 0 = all pages
includePageTexts	boolean	Also output a per-page text array — ideal for RAG chunking

Output

One item per PDF:

{
  "url": "https://arxiv.org/pdf/1706.03762",
  "fileName": "1706.03762",
  "fileSizeBytes": 2215244,
  "numPages": 15,
  "pagesExtracted": 15,
  "text": "Attention Is All You Need ...",
  "textLength": 39432,
  "wordCount": 6120,
  "title": "Attention Is All You Need",
  "author": null,
  "creator": "LaTeX with hyperref",
  "producer": "pdfTeX-1.40.25",
  "creationDate": "D:20240410010203Z",
  "modDate": null,
  "encrypted": false,
  "pageTexts": null
}

With includePageTexts: true, pageTexts becomes ["page 1 text...", "page 2 text...", ...].

Use cases

• 🤖 AI / RAG pipelines — turn PDFs into LLM-ready text; use pageTexts as natural chunk boundaries for embeddings.
• 🔎 Document search & indexing — full text plus title/author/date metadata to power search over a PDF corpus.
• 📊 Content & compliance analysis — word counts, page counts, and document properties across large batches.
• 📚 Bulk research ingestion — pull hundreds of arXiv / SSRN papers or vendor whitepapers in one run.
• 🧾 Report & invoice processing — extract the text layer from financial reports, statements, and structured invoices.

Run it from the API

Trigger the Actor and get results in one call:

curl -X POST "https://api.apify.com/v2/acts/darknezz~pdf-text-extractor/run-sync-get-dataset-items?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"pdfUrls":["https://bitcoin.org/bitcoin.pdf"],"includePageTexts":true}'

Or from Python with the Apify SDK:

from apify_client import ApifyClient

client = ApifyClient("YOUR-APIFY-TOKEN")

run = client.actor("darknezz/pdf-text-extractor").call(
    run_input={"pdfUrls": ["https://arxiv.org/pdf/1706.03762"], "includePageTexts": True}
)
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["title"], item["wordCount"], "words", "-", item["url"])

Scheduling: attach an Apify Schedule (e.g. hourly) to keep a folder of report URLs continuously ingested into a dataset that feeds your pipeline.

Pricing

Going FREE on 24 Aug 2026. Until then it runs on a small pay-per-event fee (per PDF extracted, plus Apify's standard actor-start and dataset-item micro-fees — roughly $5 per 1,000 PDFs); after the scheduled change this actor is a free tool on the Apify Store — you only pay Apify's standard platform costs for running it, not per file. No monthly minimum.

FAQ

Does it OCR scanned PDFs? No — it extracts the embedded text layer. Scanned image-only PDFs return empty text (the item still includes metadata and page counts, so you can detect and route them to an OCR step).

Password-protected PDFs? Files encrypted with an empty password (common "secured" PDFs) are extracted automatically. User-password-protected files are skipped with a warning rather than failing the run.

What's the file size limit? 100 MB per file.

Can I limit pages for very long documents? Yes — set maxPagesPerPdf to any number; 0 extracts everything.

Do I need a proxy? No. The Actor downloads PDFs directly over HTTPS — no proxy or residential IP budget required.

What if one URL is broken? That file is skipped and logged; every other PDF in the batch still returns normally.

Which engines are supported? Any server-rendered PDF served over HTTPS — arXiv, government filings, corporate reports, manuals. If a URL redirects to a login wall or an HTML page, it's logged as a failed fetch and skipped.

Limitations

• Only the embedded text layer is extracted — scanned (image-only) PDFs yield empty text and need an OCR step (see FAQ).
• Files behind interactive JS viewers (e.g. some document portals) aren't supported — use the direct .pdf URL.
• Extraction quality depends on the PDF's own text layer; poorly generated files may have garbled ordering, which is a property of the source document, not the extractor.

Enjoying the extractor? A quick review on the Apify Store helps others find it.