PDF to Text & OCR - DOCX, scanned PDF and images to clean text (RAG-ready)

Give it links to PDF, DOCX, PNG, JPG or TIFF files; get clean text per page with metadata - $0.003 per document, $0.006 per OCR page, failed files free.

PDFs that already have a text layer are read directly - no OCR, no OCR price. Only scanned pages and images go through OCR (English, German, French, Spanish, Russian, Italian, Portuguese). Optional Markdown and paragraph-aligned chunks with page numbers for search and RAG pipelines.

Try it now: press Start with the prefilled input (one public NIST PDF). It returns the full text, page by page, for $0.003.

What you get per document

• text - full text; pages[] - text per page with method (text-layer or ocr) and OCR confidence
• pages_count, pages_ocr, text_layer (full / partial / none), language, ocr_confidence, words, title, author
• optional markdown (DOCX headings, lists and tables are kept) and chunks[] for RAG (about N characters, paragraph-aligned, with page_start / page_end)
• failed documents come back as status: "error" with the reason, and are free

Input

• URLs - direct links to files (file type is detected from content)
• Key-value store keys - files you uploaded to a store (upload with the real Content-Type)
• OCR languages - English, German, French, Spanish, Russian, Italian, Portuguese (English is always included)
• OCR mode - auto (OCR only pages without text), force, never
• Markdown, chunk size and overlap, max pages per document, max file size, OCR DPI

Pricing (pay per event)

• $0.003 per document without OCR (PDF with text layer, DOCX)
• $0.006 per OCR page (scanned PDF pages and images; the pages of such a document that already have text are not charged)
• No start fee. Failed documents are not charged. Your maximum-charge setting is respected: the actor stops before it would exceed it and never returns unpaid OCR pages.

Notes and limits

• OCR quality depends on the scan. Tesseract handles clean scans well; low-resolution, skewed or handwritten pages are weaker. Check ocr_confidence.
• Password-protected PDFs are reported as failed. Images embedded inside DOCX files are not OCR'd.
• DOCX files have no fixed pages: the text comes back as one page; pages_count is filled when the file declares it.
• Only the files you provide are processed; no websites are crawled. A URL is fetched once, as a plain download, with the user agent SipkiDocsToText/1.0 (+https://apify.com/sipki/docs-to-text-ocr). If the host answers 401, 403 or 429, the document is reported as failed (free): no retry, no workaround, no proxy.
• Links to private or internal addresses (private ranges, loopback, link-local, cloud metadata such as 169.254.169.254) are refused, also after redirects.
• Process only files you have the right to process. Documents you submit may be protected by copyright; you are responsible for that.
• Open-source components: pypdfium2 (Apache-2.0/BSD-3), pdfminer.six (MIT), python-docx (MIT), Pillow (HPND), Tesseract (Apache-2.0).

How to use

1. Paste direct file links into URLs, or upload files to a key-value store and list their keys.
2. Optional: OCR languages, OCR mode (auto / force / never), Markdown, chunk size.
3. Press Start. Export as JSON, CSV or Excel, or call it from the Apify API, a schedule, an integration or the Apify MCP server.

Who uses it

• Teams building search or RAG over contracts, reports and manuals.
• Back-office automation - read invoices, forms and scanned letters into text.
• Researchers - bulk-extract text from public reports.

More questions

How much do 100 text PDFs cost? $0.30. A 10-page scanned PDF costs $0.06 (10 OCR pages). Does it read handwriting? Not reliably. Clean printed scans work best; check ocr_confidence. Are images inside DOCX files OCR'd? No, only the document text.

Not affiliated

Not affiliated with, endorsed or sponsored by NIST, the Tesseract OCR project, or the authors of the open-source components named above. Names are used only to identify them.

Content rights

You provide the documents and are responsible for having the right to process them and to use the extracted text. The Actor does not check copyright or licences of your files.

Acceptable use

This Actor is a tool. You decide which sites, companies or files it processes and what you do with the results, and you are responsible for that use. Use it only where you have the right to access the data, and in line with applicable law (including data-protection, anti-spam and copyright law) and the terms of the sources. Do not use it to send unsolicited messages, to build lists for spam, to bypass blocks or access controls, or to place excessive load on any site. If a site owner asks us to stop, we may exclude that site. Sipki, LLC does not receive or store your results; they stay in your Apify storage.

No warranty

The Actor and its output are provided "as is", without warranty of accuracy, completeness or fitness for a purpose. OCR output can contain errors. Verify results before relying on them. Liability of Sipki, LLC is limited as set out in the Apify Standard Actor Contract that governs your use of this Actor.

Personal data

Your files and the extracted text may contain personal data (for example names, ID numbers, medical or financial details in scans). For such data you are the controller and Sipki, LLC acts only as your processor under section 5 of the Apify Standard Actor Contract. Files and extracted text are processed only to produce your dataset; the Actor stores nothing outside your run's storages, and the creator does not access run data except on your support request. You need a lawful basis for your processing and must handle data-subject requests. To request deletion of anything related to you, open an Issue on this Actor's page.

For site owners (opt-out)

When you give the Actor a link, it downloads that single file with the user agent SipkiDocsToText/1.0 (+https://apify.com/sipki/docs-to-text-ocr). It does not crawl and does not read robots.txt, because it only fetches links that a user supplies. To have your domain excluded, or to report a problem, open an Issue on this Actor's page: your domain will be added to the block list within 3 business days, and every run checks that list.