Extract the Text from a PDF
Some PDFs contain text. Some contain a photograph of text. They look identical on screen and behave completely differently the moment you try to copy anything out, which is why “I can’t select the text in this PDF” is such a common and confusing problem. This page reads whatever text the document actually contains, hands it to you as plain text, and tells you outright when the answer is that there is none.
Drop a PDF to pull its text out
Reads the text layer the document already contains — no OCR, and no upload. A scanned page with no text layer will come back empty, and the page will tell you so.
Do something with the text, not just extract it
Once the words are out, Moda’s agent can turn them into a deck, a document, or a diagram rather than a text file.
Try Moda free →Text layers, and the reason scans come back empty
A PDF made by a word processor, a design tool, or a LaTeX compiler stores its words as characters with positions and a font — that is the text layer, and it is what makes the document searchable in your reader. A PDF made by a scanner or a phone camera stores a picture of a page. There are no characters in it at all; the letters you can plainly see are arrangements of pixels. Extraction reads the first kind and finds nothing in the second, and no tool can extract text that was never in the file. When every page comes back empty here, the page says so in those words rather than showing you a blank box — because a blank box reads as a broken tool, and the actual finding (“this is a scan”) is the useful information.
This page does not do OCR, on purpose
Optical character recognition is what turns a picture of text into text, and it is a genuinely different kind of operation: a machine-learning model, tens of megabytes of weights, and a result that is a guess rather than a read. A good OCR pass on a clean scan is 98% accurate, which sounds excellent until you notice it means roughly one wrong character per line, silently. Mixing the two behind one button would mean you could never tell whether what you are looking at was read or guessed. So this page reads, and says so. If you need OCR, a desktop tool or a dedicated service is the right answer, and knowing that your document needs it is exactly what this page tells you.
How the layout survives — and where it does not
Line breaks come from the document itself: a PDF marks the end of a text run, and this tool uses those marks rather than guessing from coordinates. That produces good results on prose and mixed results on layouts that were never linear. Multi-column pages extract in the order the columns were written, which is usually right and occasionally interleaves two columns. Tables lose their structure, because a PDF table is not a table — it is text positioned in a grid, with no record of which cell anything was in. Headers, footers and page numbers extract along with everything else, since nothing in the file marks them as furniture. Page markers can be switched off if you want one clean run of prose.
Where the text goes
Copy the whole thing to your clipboard, or download it as a UTF-8 `.txt` named after your document. Both come from the same extraction, so what you copy is exactly what you would have downloaded — there is no second code path that could quietly differ. Character count and page count are shown so you can sanity-check the result: a 60-page report producing 400 characters means most of it is images, which is worth knowing before you paste it somewhere.
Frequently asked questions
Is my PDF uploaded to extract the text?
No. The document is parsed by a PDF interpreter compiled to WebAssembly that runs inside this page, so the file never leaves your machine and the text never touches a server. That is the property that makes this usable on a contract or a medical record, which is most of what people extract text from.
Why is the text jumbled or in the wrong order?
Almost always a multi-column layout, or a document whose text runs were written out of visual order — a PDF has no obligation to store its text in reading order, and some generators do not. Nothing can fully fix that without inferring layout from coordinates, which trades one class of error for another. If the document is a form or a table, expect structure to be lost.
Some pages have text and some are empty. Why?
A mixed document: the digitally produced pages have a text layer and the scanned or image-only pages do not. This is extremely common in signed contracts — the body was exported from a word processor and the signature page was scanned back in. The page tells you how many came back empty.
Can I extract text from a password-protected PDF?
No, and it is refused by name rather than partially processed. Removing PDF passwords is deliberately not something these tools do. Open the file with its password in a reader and save an unprotected copy.
Does it keep bold, headings, or bullet points?
No — the output is plain text. Styling in a PDF is a property of the drawing instructions, not of the characters, so there is no faithful way to turn it into markup without guessing. Bullet characters usually survive because they are real characters in the text; the indentation around them may not.