Skip to content
RemixPDF

Convert a PDF, free — to text, Markdown, HTML or images

Pick what you need it as. Plain text and images come out exact; Markdown and HTML keep the structure — headings, lists, links, pictures — with anything lost named in the file. Word export is coming as a labelled server feature.

  • Turn it into text, Markdown, HTML or images.
  • Free
  • No signup
  • No watermark

Runs on your device

What you have to hand while you work

Why it works

Every format is labelled for what it keeps.

Images and plain text come out exact — a page really is a picture, and text really is text. Markdown and HTML keep the structure — headings, lists, links and images — and name whatever did not survive. Word export, when it arrives, will be labelled approximate, because calling it anything else would be a lie you discover after the deadline.

It says why Word is hard.

A PDF describes where each letter goes on a page. A Word document describes paragraphs that reflow. Converting between them means guessing, and the guess shows most in tables, columns and anything hand-positioned.

Two better routes when the guess is not good enough.

If you only need the words, plain text is exact. If you need to change something in place, editing the PDF directly avoids the round trip entirely — and we will suggest that rather than sell you a conversion.

Most conversion tools in this category make the same promise and break it in the same place. You upload a PDF, you get a Word document, and it looks nearly right until you scroll to the page with the table on it — at which point the columns have collapsed, a heading has drifted into the middle of a paragraph, and half an hour of tidying is ahead of you.

That happens for a structural reason rather than a quality one. A PDF is a set of instructions about where marks go on a page: this glyph at this coordinate, in this font, at this size. A Word document is a description of content that reflows: this is a heading, this is a paragraph, this is a table with four columns. Converting a PDF to Word means inferring all of that structure from positions alone, because the structure was never recorded in the first place. It is genuinely a guess, and the guess is good for simple prose and poor for anything laid out by hand.

So the formats on the counter are the honest ones. Images are exact — a page is already a picture, so turning it into one loses nothing. Plain text is exact for what it claims: every word that was in the file, with no layout at all. Markdown and HTML sit in between — they carry the structure a reader can rely on (headings, paragraphs, lists, links, images) and say plainly what they dropped, including the tables whose columns could not be trusted to a guess. Word export is on its way as a labelled server feature — and only ever described as approximate, with the reason stated and a specific warning when the document is full of the things that break.

And when a conversion is the wrong answer, this says so. If the reason you want a Word file is to change one sentence, editing the PDF directly does that without the round trip and without the layout damage — which is a worse outcome for our conversion numbers and a better one for you.

Scanned documents are the other honest case. A scan contains no text, only a photograph of it, so converting one to text would hand you an empty file. The tool says that, and points at making the document searchable first, rather than letting an empty download look like a success.

Today’s formats run entirely in this browser — nothing is uploaded, your original is untouched, and there is no signup, no queue and no cap.

How it goes

  1. Open the PDF.

    It opens here, and the tool reads what kind of document it is.

  2. Choose the format.

    Plain text, Markdown, a self-contained HTML page, or a picture of each page. Word, Excel and PowerPoint are honestly marked as not built, not offered as if they were ready.

  3. Read the note if there is one.

    A document that will not convert cleanly gets a plain explanation rather than a broken file.

  4. Download it.

    Your PDF is untouched — the converted file is new.

What you can check for yourself

Watch the network while you work — nothing leaves this device on the default path. Then go offline and keep working — it all still runs. The two-minute checks live at how your files are handled.

Frequently asked questions

Word export is on the way as a labelled server feature — and no, it will never look exactly like the PDF, whatever anyone promises. Paragraphs, headings and plain text usually come through well. Tables, multi-column layouts and anything positioned by hand are where it goes wrong, because those have no equivalent in a reflowing document.

Because the two formats describe different things. A PDF says "this letter goes here"; Word says "this is a paragraph". Going backwards means inferring structure that was never recorded, so it is always a guess.

Plain text, Markdown, HTML, PNG and JPG — all running in this browser. Text and images are exact. Markdown and HTML keep the structure — headings, lists, links and images — but not the layout, and what was lost is said in the file itself.

The text is kept, in reading order, inside a fenced block — with a note saying the columns were not preserved. Guessing at column boundaries produces tables that look right and are wrong, so we keep the words and say so rather than emit a false table.

Not yet. Table extraction is a harder problem than it looks, and doing it badly produces a spreadsheet worse than useless, so we would rather not offer it until it works.

There is no text to convert, only a picture of it. Making it searchable first turns it into a real conversion; we will say so rather than handing you an empty file.

Usually not. Editing the words in the PDF directly avoids the conversion and the trip back, and it keeps the layout exactly as it is.

Related tools

Ready? Drop your PDF above and it opens right here.

Choose a PDF

Language

English and Spanish are fully translated. The rest are listed because they are coming — pick one and the page stays in English rather than showing you half a translation.