Unlimited OCR for PDFs, Scans and Photos
Upload a PDF, JPG or PNG file and get Markdown or plain text back.
Free trial: 1 document a day (up to 3 pages, 10 MB), no account needed. The exact model is shown on every result. Paid plans coming soon. The free trial currently runs the Unlimited-OCR 2 model.
- Free trial: 1 document a day (up to 3 pages, 10 MB), no account needed.
What Is Unlimited OCR?
"Unlimited" is part of the model's name. Its developers designed the Unlimited OCR model for long documents. It does not mean LongDoc OCR offers unlimited use: the free trial has daily and per-file page limits. Free trial: 1 document a day (up to 3 pages, 10 MB), no account needed. The exact model is shown on every result. Paid plans coming soon.
Looking for the Unlimited OCR 2 model itself? See the Unlimited OCR 2 model explained.
Which Model Reads Your File?
The free trial currently runs the Unlimited-OCR 2 model. The exact model is shown on every result, with a link to how we process files.
What the Free Trial Extracts
Results from small tests, not a guarantee.
Printed English
About 99% character accuracy on the pages we tested. One phone photo scored lower and missed a line, so always check your result.
Printed Traditional Chinese
About 97% in our test, with a few wrong characters. Text stays in Traditional characters and is not converted to Simplified.
Two-column pages
Two-column pages come back in the right reading order.
Tables
Rows in scanned tables and in tables that continue across pages were all captured in our tests. A table that spans pages may come back as several separate tables. Merged cells in scanned tables were recognized correctly, but merged header cells in PDFs with a text layer may be lost.
Clean Markdown
Page headers, footers and page numbers are left out of the Markdown.
How to Use It
- Upload — Choose a PDF, JPG or PNG file.
- Tick the consent box — Agree that your file is processed by a third-party OCR provider on servers in mainland China.
- Extract — Press Extract text and wait for the result.
- Copy or download — Copy the result, or download it as a Markdown or TXT file.
What It Can't Do (Yet)
Not promised yet
Handwriting results are mixed, and we haven't finished testing long documents or spreadsheet export, so we don't promise them.
Proofread
Proofread names and numbers, and check that no lines are missing.
Pricing
Paid plans coming soon.
Privacy
At this stage, every file is processed by a third-party OCR provider on servers in mainland China, and you agree to this before you upload. We delete our copy of your original file right after processing and keep our copy of the result for 24 hours. We copy each result to our own storage as soon as it is ready and never hand you the provider's link. The provider's documentation says result links can last up to 30 days, and how long it keeps original files is governed by its own terms, which we cannot control. See who processes your files.
Who Uses It
Researchers
Scanned papers and handouts turned into editable Markdown.
Two-column papers and scans
Two-column pages and scanned material turned into text in reading order.
RAG data prep
Markdown for retrieval and LLM pipelines. Choosing a model to self-host instead? See how Unlimited OCR compares with DeepSeek-OCR.
Unlimited OCR FAQ
Is LongDoc OCR really unlimited?
No. "Unlimited" is part of the model's name: its developers designed the Unlimited OCR model for long documents. It does not mean LongDoc OCR offers unlimited use. The free trial gives you 1 document a day (up to 3 pages, 10 MB) with no account needed. Paid plans are coming soon.
Which OCR model processes my file?
The exact model is shown on every result, with a link to our Privacy Policy, which explains who processes your files.
Is LongDoc OCR an official site, or related to other sites with a similar name?
No. LongDoc OCR is an independent service and is not affiliated with or endorsed by Baidu. We also have no connection to any other website or tool with the same or a similar name. Our only address is unlimited-ocr2.com. To learn about the open-source model itself, see our Unlimited OCR 2 guide.
What files can I upload?
PDF, JPG and PNG files, including scanned PDFs with no text layer and photos taken with a phone. Each file can have up to 3 pages and 10 MB.
Does it work on Chinese text?
In our small test, printed Traditional Chinese was read with about 97% character accuracy, and the text stayed in Traditional characters rather than being converted to Simplified. A few characters were wrong, so always proofread. We haven't tested Simplified Chinese or Chinese handwriting yet.
Does it handle two-column pages and tables?
In our tests, two-column pages came back in the right reading order, and page headers, footers and page numbers were left out of the Markdown. All rows of a scanned table and of a table that continued across pages were captured, but a table that spans pages may come back as several separate tables. Merged cells in scanned tables were recognized correctly; in PDFs with a text layer, merged header cells may be lost. These are results from small tests, not a guarantee. We haven't finished testing long documents.
Can I buy a paid plan?
Not yet. Paid plans are coming soon. For now, you can use the free trial.
How long do you keep my files?
For our own copies, we delete your original file right after processing and keep the result for 24 hours. Our OCR provider keeps its own copies under its own terms, which we cannot control. Its documentation says result links can last up to 30 days (in our tests they expired after about 3 days). We copy your result to our own storage right away and never give you the provider's link. How long the provider keeps original files is governed by its terms. Details are in our Privacy Policy.
Who processes my files, and where?
At this stage, every file is processed by a third-party OCR provider on servers in mainland China. Before you upload, we ask you to agree to this. Our Privacy Policy names the provider.
Are my files used to train AI models?
LongDoc OCR does not use your files or results to train our own models. How our third-party OCR provider handles your data, including whether it is used for model training, is governed by its own terms, as described in the "Who processes your files" section of our Privacy Policy.
What happens if OCR fails or free capacity runs out?
If OCR fails, your free use is restored, so the failed attempt does not count toward your daily documents. You'll see "OCR failed and your free use has been restored. Please try again later." If the free capacity for the whole site is used up for the day, you'll see "Free capacity for today is used up, please come back tomorrow." Free limits reset at 00:00 UTC.