LongDoc OCR

PDF to Markdown: Convert PDFs and Scans with OCR

LongDoc OCR turns a PDF, JPG or PNG into Markdown in your browser: upload the file and get headings, paragraphs and detected tables as Markdown you can copy or download as .md or .txt. It uses OCR, so scanned PDFs work too, and you don't need a GPU. Try 1 page free without an account; a new account gets 10 free pages once.

See pricing for long PDFs

Choose a PDF, JPG or PNG file first.Cost: 1 credit per page

The exact model is shown on every result. Credit packs and paid plans are available for more pages. The free trial currently runs the Unlimited-OCR 2 model.

How do I convert a PDF to Markdown?

Upload the file, let OCR run, then copy or download the Markdown.

  1. Choose a PDF, JPG or PNG (free: up to 1 page and 10 MB; paid: up to 500 pages and 40 MB).
  2. Agree to where the file is processed, then click Extract text.
  3. Check the model name shown on the result, proofread, then copy it or download it as .md or .txt.

On paid plans each page uses 1 credit; credits are returned automatically if processing fails, times out or is cancelled.

Which PDFs convert well to Markdown?

Printed, text-heavy documents convert best: reports, papers, manuals, contracts and book pages, scanned or digital. In our small test (5 documents, 8 pages), two-column pages came back in the right reading order, and page headers, footers and page numbers were left out of the Markdown. Phone photos work but were less accurate, and handwriting is not a supported use.

What does the Markdown output look like?

The output is plain Markdown: # headings, paragraphs in reading order, and detected tables as Markdown tables. The block below shows the format only; it is not a real test result:

# [Document title]

## [Section heading]

[Paragraph text in reading order …]

| [Column A] | [Column B] |
|---|---|
| [cell] | [cell] |

Headings are detected, but heading levels were not always consistent in our small test, so check them before you publish or pass the file to another tool.

Can I convert a long PDF to Markdown?

Yes, with a credit pack or monthly plan: each file can have up to 500 pages and 40 MB, at 1 credit per page. Our largest end-to-end test so far is a 68-page PDF converted in a single upload (small sample, not a guarantee). The free trial is 1 page per document, so use it to check quality, not for long PDFs. See pricing.

Where is my PDF processed, and how long is it kept?

Free pages without an account are processed by a third-party OCR provider (named in our Privacy Policy) on servers in mainland China. The 10 free sign-up pages and paid files are processed on our own GPU server in China. While that GPU server is offline, files of up to 5 pages may go to a third-party OCR provider (named in our Privacy Policy) instead: you agree to this once, by ticking the box before you upload, and you are not asked again when it happens. After the 10 sign-up pages are used, free pages also go to a third-party OCR provider (named in our Privacy Policy). Results are kept for 24 hours (free) or 72 hours (paid), then deleted. LongDoc OCR does not use your files or results to train models. Details: Privacy Policy.

What can't it do yet?

It returns Markdown or plain text, not a searchable PDF or Word file. Simplified Chinese, LaTeX formulas, code blocks and CSV/Excel export are not tested and not promised. Long tables may be split into several tables, merged header cells in PDFs with a text layer may be lost, and there is no speed promise. Always proofread OCR output.

How does this compare with running the model yourself or using a text-layer converter?

Each option fits a different need:

  • Running the open-source model yourself (from its GitHub / Hugging Face release) gives you full control and keeps files on your own machine, but needs a GPU and setup. See what Unlimited OCR 2 is and how to run it.
  • Text-layer PDF converters extract text already stored in a digital PDF, which can be fast and exact for those files, but they cannot read scanned pages that have no text layer.
  • LongDoc OCR runs OCR from your browser with no setup and reads scans and photos; files are processed in China, and paid use is charged per page.

Is it free, and is "Unlimited" really unlimited?

It is free within daily limits, and "Unlimited" is a model name, not a usage promise. Without an account: 1 document a day per network, up to 1 page and 10 MB. New accounts: 10 free pages once. Each network gets at most 3 free documents a day. For more, buy a credit pack (from $4.99 for 250 credits that never expire) or a monthly plan, and cancel any time.

LongDoc OCR: PDF to Markdown at a glance

  • Input: PDF, JPG, PNG. Output: Markdown or plain text (.md / .txt download).
  • Free: 1 document a day without an account, up to 1 page and 10 MB; new accounts get 10 free pages once.
  • Paid: up to 500 pages and 40 MB per file; 1 credit per page; credit packs from $4.99 (250 credits).
  • Largest end-to-end test so far: a 68-page PDF in one upload (small sample, not a guarantee).
  • Tables: detected; structure not guaranteed.
  • Processing: a third-party OCR provider (named in our Privacy Policy) in mainland China (free pages without an account, and free pages after the 10 sign-up pages are used); our own GPU server in China (sign-up pages and paid files); while that server is offline, files of up to 5 pages may go to a third-party OCR provider (named in our Privacy Policy) (agreed once before upload, not asked again).
  • Results kept 24 hours (free) / 72 hours (paid), then deleted.
  • LongDoc OCR is an independent service; "Unlimited" is part of the open-source model's name.

PDF to Markdown FAQ

Can I convert a PDF to Markdown for free?

Yes, within limits: without an account you can convert 1 page a day (up to 10 MB), and a new account gets 10 free pages once. After that, credit packs and monthly plans cost 1 credit per page.

Does it work on scanned PDFs?

Yes. LongDoc OCR reads each page as an image with OCR, so scanned PDFs and photos of pages work, not only PDFs with a text layer. Always proofread the result.

How long a PDF can I convert to Markdown?

Paid files can have up to 500 pages and 40 MB. Our largest end-to-end test so far is a 68-page PDF converted in one upload (small sample, not a guarantee). The free trial is 1 page per document.

Are tables kept in the Markdown?

Tables are detected and written into the Markdown output, but table structure is not guaranteed. In our small test, a long table that ran across pages came back as 3 separate tables, and merged header cells in PDFs with a text layer may be lost.

Where are my files processed and how long are they kept?

Free pages without an account are processed by a third-party OCR provider (named in our Privacy Policy) on servers in mainland China. The 10 free sign-up pages and paid files are processed on our own GPU server in China. While that GPU server is offline, files of up to 5 pages may go to a third-party OCR provider (named in our Privacy Policy) instead: you agree to this once, by ticking the box before you upload, and you are not asked again when it happens. After the 10 sign-up pages are used, free pages also go to a third-party OCR provider (named in our Privacy Policy). Results are kept for 24 hours (free) or 72 hours (paid), then deleted.

Does it mean unlimited conversions?

No. "Unlimited" is part of the name of the open-source model, Unlimited OCR 2. LongDoc OCR has daily free limits and paid page credits.

Is LongDoc OCR the official Unlimited OCR 2 site?

No. LongDoc OCR is an independent service that runs the open-source model.

Can I cancel a monthly plan?

Yes. Cancel any time with one click on your account page or by emailing support@unlimited-ocr2.com. The plan keeps working until the end of the period you paid for.