PDF to Text
The words you need, out of the PDF. Read them beside the original, put the line breaks right, and take them anywhere.
Open a PDF. Check the words. Make them useful again.
PDF to Text is in your dock. Everything you started is still there.
Keep the words.
Your document on the left. Its words on the right. Ready to use again.
A PDF goes in. Your words come out.
Reports, research or a scan. Open one PDF up to 100 MB.
OCR reads pages without text. Check its spelling against the original.
Edits stay with each page in this tab. Line settings apply to unedited pages. Save before closing.
PDFs and extracted text stay in this tab.Only display and export settings save to your account.Saving details
From a fixed page to words you can use
Open a PDF
Choose a PDF or drop it anywhere on the page. The sample includes two text pages and a scan.
Review the words
Compare each page with its editable text. Choose your pages, tidy line breaks or read scans with OCR.
Keep what you need
Copy the selected text or save TXT or Markdown. Keep one document or a separate file for every selected page.
Using PDF to Text
Open one PDF. Embedded text appears automatically, without waiting for an OCR model. Your original file stays unchanged.
Pick the pages
Leave Pages to export blank for every page. Enter a range such as 1, 3-5 and press Apply to keep those pages in your output.
The numbers refer to positions in the PDF, starting at 1. They may differ from printed page numbers. An invalid range leaves your last valid selection intact.
Shape the text
| Control | What changes |
|---|---|
| Keep line breaks | Keeps the line boundaries found in the PDF. Useful for lists, addresses and short labels. |
| Join lines into paragraphs | Joins single line breaks with spaces. Existing blank lines still separate paragraphs. PDF paragraph boundaries are not always available. |
| Join words split by a hyphen | Joins a letter, a hyphen and a line break when the next line starts with a lowercase letter. Check genuine compound words afterwards. |
Line settings affect unedited pages. Once you edit a page, its text stays exactly as you entered it. Restore extraction discards that page’s edits and applies the current settings again.
Review beside the original
Use Previous page, Next page or Go to page to move through the document. Your edits stay with each page.
Compare shows the original page beside the transcript. Text only gives the words more room. The word and character counts describe the current transcript; the selection count describes your export.
Narrow, Wide or Full at the top right opens the tool in a focused workspace. Minimise keeps your PDF in the dock. Close (×) or Escape returns to the page.
Read scans with OCR
The tool flags selected pages without embedded text. Choose English, French, German, Spanish, Italian or Portuguese, then press Read scans with OCR.
- OCR reads selected pages that have no embedded text. It leaves existing text and your edits alone.
- The first run loads an engine and language model from
models.floi.dev. The page images stay in your browser. - A blank page may still have no text after OCR. Recognition confidence is an engine estimate, not a guarantee of correctness.
- Stop finishes the current OCR page and keeps completed pages. Read the remaining scans again when you are ready.
Copy or save the result
| Control | What you receive |
|---|---|
| Copy text | All selected pages, using your current format and page-number setting. |
| Save text | One UTF-8 TXT or Markdown file, in PDF page order. |
| Save pages | One text file per selected page. The shared save menu offers one ZIP or separate files. More than 50 files always use a ZIP. |
| Include page numbers | Adds plain-text separators such as --- Page 3 ---. Markdown always includes page headings. |
| Open in / Share | Passes the combined result to another tool or shares it through the file bar. Share uses the selected-page batch. |
| Open another / Clear | Opens a different PDF or returns to the empty panel. Both discard this tab’s current transcript. |
Save before closing or reloading. PDFs, previews and transcript edits stay in tab memory. Only your line, format and language settings save automatically to a signed-in account.
What it will not do
This extracts words. It does not recreate a Word document or guarantee the reading order of columns and tables.
- One PDF at a time, up to 100 MB and 2,000 pages. Extracted text and edits are limited to 10 million characters.
- OCR reads up to 100 selected pages per run. Each page is rendered within 8 megapixels and 8,192 pixels per side.
- No forced OCR of an existing text layer, handwriting guarantee, table-cell reconstruction or recovery of missing font mappings.
- Large PDFs depend on your device’s memory. Stop during extraction keeps only pages already read. Reopen to read the full document.
- To keep the PDF appearance and add searchable text, use Searchable PDF instead.
A sample you can check
Try a sample opens a synthetic three-page PDF. The first two pages contain embedded text; the third is a scan. The first includes docu- at a line ending and ment on the next line. Use it to inspect cleanup and OCR before opening your own document.
Frequently asked questions
Can I convert a scanned PDF to text?
Yes. Pages without embedded text are flagged. Choose the scan language and press Read scans with OCR. Recognition runs on your device. Check the result, especially names and numbers.
Does this send my PDF or extracted text to a server?
No. The PDF, its previews, passwords and extracted text stay in this tab. OCR loads its engine and language model from floi’s model host. It does not send your pages there. Only display and export settings can save to a signed-in account.
Why is the copied text in the wrong order?
A PDF can position each word separately without describing paragraphs or columns. The extractor follows the text sequence in the file. Compare it with the original and correct the text before saving. Complex columns and tables need particular care.
Will PDF to TXT preserve tables and formatting?
TXT keeps characters and line breaks, without fonts, pictures or table cells. Markdown adds a heading for each page. It does not reconstruct document headings, links or tables.
Can I extract text from a password-protected PDF?
Enter the opening password when prompted. A wrong password lets you try again. The password is cleared after submission and never saved. You must have permission to use the document.
Something missing? Request a feature