Loading Techie Utils…

Online PDF Editor

Move and retype what is already on the page, reorder, annotate, stamp and sign, with the document in front of you. Free, and runs in your browser.

About PDF Editor

A ribbon of tools across the top, pages down one side, the page you are working on in the middle, and the properties of whatever is selected on the right. The ribbon is grouped by what you came to do — View, Annotate, Shapes, Pages — because "draw a rectangle" and "make that rectangle dashed" are different questions, and a single column of tools has to answer both at once, which means it is always mostly showing controls for something nobody is doing.

The expand button gives the editor the whole window. That is not a luxury here: a PDF page is a tall thing being looked at inside a page that is itself scrolling, inside a site with a navigation rail down one side, which at laptop height leaves the document about a third of the screen. Expanded, the site's chrome goes and a small mark takes its place in the corner; Escape brings it back.

Click a page to open it; drag it somewhere else, turn it upright, delete it, duplicate it, insert a blank one, or bring in pages from a second file. A blank page arrives the size of the page you are on, or at A4, Letter, Legal or A3 either way up — and it is an ordinary page from the moment it appears, so you can write on it, stamp it, sign it, turn it and move it exactly as you can any other. Turn on the strip of page pictures across the top when you want to *see* the document rather than navigate it — it is off by default because it costs a band of the height the page itself is drawn in, and that trade is only worth making while you are looking for something.

The catalog already has a merger, a splitter, a page deleter, an organiser and a rotator. Each does its one job correctly, and none of them lets you see the document while you work. Reordering pages by typing "5,2,9,1" into a field is an operation carried out blind — it is not obvious it went wrong until the file is opened afterwards, which is why the organiser is the tool people end up running three times.

The editing model is worth knowing because it decides what this can do well. The document is held as a *list of references* to pages in the files you opened, not as a PDF that gets rewritten after every click. So reordering, rotating and deleting are free, undo is free, and the bytes are written exactly once at the end from the original files. A page that was moved, turned and moved again has still been through one writer rather than three — which matters, because every pass through a PDF writer is another chance to drop something it did not understand.

Any font, not just the three a PDF reader is born with. Helvetica, Times and Courier are the only faces that can be used without putting the font itself in the file, which is why every browser PDF editor offers three families and stops. This one will take a real font two ways: Chrome and Edge can hand over everything installed on the machine once you allow it, and every browser can open a .ttf or .otf file. Either is written into the saved document, cut down to the characters you actually used — so it looks the same on any machine, and it can set the curly quotes, dashes and non-Latin characters the built-in faces cannot.

That is also what makes editing a line look like editing rather than like replacing: with the machine's fonts loaded, the original font's name is read out of the document — subset tag and all — matched against them, and the replacement is set in its own face, in the right weight and slant. Without them it falls back to the nearest of the three, which is the old behaviour and is visibly a different typeface on anything set in Calibri or Garamond.

A running header and footer, three slots along the top and three along the foot, with {page}, {pages}, {date} and {name} filled in per page — so "Page 3 of 12" counts itself. Put it on every page, on the ones you have selected, or on the one in front of you. What you get is ordinary text on each page, which is the point: afterwards you can move it, recolour it, retype it, or delete the one on the page that should not have had it. Adding a header and editing a header are the same feature.

The text already on the page can be edited. Pick up the text tool and every line the document actually holds is outlined on the page; click one and it becomes a field with the original words in it, at the original size, in a matching face, in the right colour — including a white heading on a blue banner, because the paper and the ink are read off the page as it has been drawn rather than assumed to be black on white. There is an *edit all* for taking a whole page at once. If the outlines never appear, the page is a scan: pictures of words rather than words, and it needs OCR before anything can read it, which is a genuinely useful thing to learn in one click.

Changing text needs its limit stated rather than discovered. A PDF does not hold editable text. It holds glyphs at co-ordinates in the font's own encoding, and the words you read are emergent — retyping one means re-measuring the run, re-flowing everything after it, and re-embedding a font subset that may not carry the new characters. So this does what every browser-based PDF editor does: it covers the old text with a filled box and writes new text on top. That genuinely changes what the page shows. It does not change the text objects underneath, so anything that reads the file rather than looking at it still finds the original words — which is why a real redaction has to rasterise the page, and the Compress PDF tool's redraw mode is what does that.

Pictures go on the same way text does. Choose any file the browser can open — a phone's HEIC, an SVG, a screenshot — and it is placed on the page, centred and sized to fit, ready to be dragged. It is converted once on the way in, because a PDF holds PNG and JPEG and nothing else, and the format is chosen from the picture itself: anything with transparency has to be PNG, and a photograph as PNG is three times the size for no visible difference. The same picture placed twice is embedded once, so a letterhead on forty pages costs one copy of the image.

Shapes are the other half of marking a document up: rectangles with adjustable corners, ellipses, triangles, diamonds, stars, lines, arrows, and polygons and polylines you build a corner at a time. Each carries its own line colour, fill, width, dash pattern and opacity, and a line can have a head at either end or both. An arrow points at the end of the drag you made rather than at a corner of its box, which sounds obvious and is the thing most annotators get wrong. A PDF has drawing operators for rectangles, ellipses and lines and for nothing else, so everything past those is written as the path it really is.

Marking up text has its own tools beside those: highlight, underline, strike through, and a squiggly underline — each a drag across the words rather than a box around them, so they work on a page this editor cannot read as text at all. Freehand drawing is there too, and a pen below full opacity and a few points wide is a highlighter that follows the words instead of boxing them. An eraser takes your own marks off again, one click or one sweep at a time; the page underneath cannot be rubbed out, because it is the file you opened.

A sticky note is the one mark that is not ink on the page. It becomes a real PDF comment — an annotation object with your text and your name in it — so Acrobat, Preview and Firefox all show it as a note that opens and can be replied to, it stays out of the way of the page, and printing the document does not print it. Nothing else here can say something *about* a page rather than on it.

Everything you place can be resized and turned. Select a mark and eight grips appear round it with a ninth above on a stem: drag a corner to resize, drag the one above to turn, and hold shift while turning to snap to fifteen degrees. A corner drag on a picture keeps its proportions; on text it changes how big the words are, because text has no box of its own. The grips work whatever tool is in hand, so drawing three rectangles and then making the third one bigger is not a trip back to the select tool. Rotation is the piece worth knowing is correct: the preview turns a mark about its middle with CSS, while a PDF turns a shape about its bottom-left corner with the y-axis the other way up — the two are reconciled in one place, and checked against the bytes that actually get written.

Every object on the page carries its own toolbar, floating just above it: turn it a quarter each way, send it forward or back in the stack, flip a picture across or over, duplicate it, put a copy of it on every page, remove it. Stacking order is the order things are drawn, which is what decides whether a logo sits on the white box or under it. "On every page" is how a letterhead, a watermark or an approval stamp gets applied to a forty-page document without forty placements. Below the page are zoom, a page number you can type into, and fit-width, fit-page and actual-size — so a 5pt footnote can be worked on at 400% and a whole page checked at once.

Signing is three different things people mean by one word, and all three are here. Draw your signature on a pad with a pointer or a trackpad; type your name and have it set in the machine's own handwriting face; or upload the photograph of the one already on paper. Whichever you use, it can be kept — in this browser profile, never uploaded — so signing the second document is one click rather than a second attempt at drawing your own name with a mouse. A drawn signature is placed as vector ink, which stays crisp at any size and costs a few hundred bytes; a typed or uploaded one is a picture, because none of the fourteen faces a PDF reader must have looks remotely like handwriting.

Be clear about what all three are: they draw a *picture* of a signature. Nothing is cryptographically bound to the document, and anybody can copy the image onto another one. That is equally true of every scanned signature in every contract, which is why it is still useful and why it is not called a digital signature here. Stamps work the same way and are offered as presets — today's date, Approved, Draft, Confidential, Reviewed by — each one an ordinary text mark, so it can be recoloured, resized and turned diagonally across the page like everything else.

The one action that cannot be undone asks first. Undo, redo and "throw everything away" used to be three unlabelled circular arrows in a row, and only the third was irreversible — reaching for redo and hitting it instead is not carelessness, it is what that arrangement asks for. Revert is now separated, named with a word rather than a third arrow, and it tells you how many changes are at stake before it does anything.

Undo is the quiet part that decides whether any of this is usable. A keystroke is genuinely a change, so recording one step per keystroke is literally correct and completely useless — undoing a sentence becomes forty presses, and the history fills up with the letters of one word. So a run of related changes close together collapses into a single step: typing a caption, dragging an opacity slider, holding the arrow key on a number field. Each is one undo. Touching a different property, or the same one after a pause, starts a fresh step, so nothing you meant to keep gets swallowed by the next thing you did.

The keyboard does the work the pointer is bad at. Undo and redo answer to all three spellings people press, the arrow keys nudge the selected mark a point at a time — ten with shift — and move between pages when nothing is selected, Delete removes the selected mark, and Escape backs out of whatever is in progress. None of them fire while you are typing into a field, which is the bug that would otherwise make a caption impossible to correct.

Nothing leaves the tab. The document is read, drawn and written here, which for PDFs is usually the entire point: contracts, statements, payslips and scans are exactly the documents nobody should be posting to a website to have their pages swapped around.

Where it helps

  • Shows the pages, and shows the one you are working on full size. Every other page tool here asks you to describe them.
  • Write text onto a page, cover something up, highlight a line, draw freehand, place a picture, stamp it or sign it — each mark movable, resizable, turnable, editable and undoable.
  • Any font, not just the built-in three: load the machine's own on Chrome and Edge, or open a .ttf or .otf anywhere. It is embedded, subsetted, and lifts the Latin-1 limit.
  • Edits the text the page already carries: click a line and type over it, with the size, face, paper colour and ink colour all taken from the page as drawn — in the original's own font when that font is loaded.
  • A running header and footer with page numbers and the date, which land as ordinary text you can then move, retype or remove page by page.
  • Says plainly when a page is a scan — no lines to edit means pictures of words, and OCR is the next step rather than a mystery.
  • Rectangles with rounded corners, ellipses, triangles, diamonds, stars, lines, arrows, polygons and polylines, each with its own line colour, fill, width, dash pattern and opacity — and an arrow that points where the drag ended rather than at a corner of its box.
  • Highlight, underline, strike through and squiggly underline as drags across the words, so they work even on a page whose text cannot be read.
  • Sticky notes that become real PDF comments — they open in any reader, carry your name, and do not print.
  • An eraser for your own marks, and text boxes with a panel and an outline behind the words.
  • One undo per thing you did, not per keystroke: typing and slider drags collapse into a single step.
  • One surface rather than four panels, so the tool that starts an action and the setting that adjusts it are not separated by three borders.
  • Fills the window on demand, which on a laptop roughly triples the height the document is drawn at.
  • Inserts blank pages — the size of the page you are on, or A4, Letter, Legal or A3 either way up — and they take marks like any other page.
  • Eight resize grips and a turn grip on whatever is selected, working under every tool, with shift snapping a turn to fifteen degrees.
  • Bold, italic, underline, strikethrough, alignment and line spacing on a text mark, with the preview showing the cut that will actually be written rather than a browser's approximation of it.
  • A ribbon grouped by what you came to do, and a properties panel that is entirely about whatever is selected — rather than one column trying to be both.
  • A strip of page pictures across the top on demand, for finding a page rather than moving one.
  • Places pictures from anything the browser can open, HEIC and SVG included, converted once on the way in and embedded once however many times it is used.
  • Stacking, flipping and duplicating from a toolbar that floats over the object rather than sitting across the room in a panel.
  • Zoom from 50% to 400%, fit-width, fit-page and actual-size, and a page number you can type into, so small print can be worked on at the size it needs.
  • Sign by drawing, by typing your name in a handwriting face, or by uploading a photograph of a real one — and keep any of them in this browser to reuse on the next document.
  • Put any mark on every page in one action, which is what a letterhead, a watermark or an approval stamp actually needs.
  • Stamps for the words people actually stamp: today's date, Approved, Draft, Confidential, Reviewed by.
  • Keyboard throughout: undo and redo, arrow keys to nudge by a point, Delete, Escape — and none of them while you are typing.
  • Cover-and-replace in one action, because doing it in two is where the old words get left showing underneath.
  • Reorder by dragging, including several non-adjacent pages at once, which lands them together.
  • Rotate, delete, duplicate and keep-only, all undoable.
  • Pulls pages in from other PDFs, so merging and reordering are the same operation.
  • Writes the file once, from the originals, so repeated edits do not compound losses.
  • Reads, draws and writes entirely in this tab — the document is never uploaded.

Limits

  • Text is written over the page, never into it — including when you edit a line that was already there. The original words stay in the file for anything that reads rather than renders it, so neither a cover box nor an edited line is a redaction.
  • An edited line is covered by a rectangle in one flat colour. Over a gradient, a photograph or a rule through the words, that rectangle can be seen — the editor says so when it happens, but it cannot avoid it.
  • The face of an edited line is a guess from what the original font was called, and the width of the new text is whatever Helvetica, Times or Courier makes it. A line that has to fit an exact column may need its size adjusting by hand.
  • A sticky note is a comment rather than content: a reader that ignores annotations, or a print, will not show it.
  • The three built-in faces carry Latin-1 only, so a character outside it is folded to the nearest one. Load a real font and that limit goes, at the cost of the font's own weight in the file.
  • Listing the machine's fonts is a Chrome and Edge feature and needs your permission. Firefox and Safari have no equivalent, so there it is font files only.
  • A loaded font lives in this tab for the session. Reopening the tool means loading it again, and a document saved with one carries the font but not the ability to keep editing in it.
  • Page images stop at the first 200 pages. Everything past that still reorders and still saves; it just shows a number instead of a picture.
  • A password-protected PDF has to have the password removed first.
  • A signature here is a picture, however it was made, not a digital signature. Nothing binds it to the document and anyone can copy it onto another one — the same as a scan.
  • Kept signatures live in this browser profile. Clearing site data loses them, and they do not follow you to another machine — which is the same fact as "they are never uploaded", seen from the other side.
  • A typed signature is drawn with whatever handwriting face this machine happens to have, so the same name can look different on a colleague's screen. It is embedded as a picture, so the document itself does not change.
  • Files stop at 64 MB, because the document, its page images and its history are all held in memory at once.
  • Bookmarks, links between pages and attachments do not survive being rebuilt. A document whose table of contents is clickable will lose that.

Frequently asked questions

Is my document uploaded?

No. It is read in the tab, drawn in the tab with the same renderer Firefox uses, and written in the tab. Nothing is sent anywhere, which is confirmable by loading the page, disconnecting from the network and using it regardless.

How do I move several pages at once?

Click each one, or shift-click for a run, then drag any of them. They keep their order among themselves and land together at the gap you drop them on — including when they started out in different parts of the document.

Why does my table of contents stop working after editing?

Because the file is rebuilt from the pages you kept, and the links pointed at page numbers in the old document. Bookmarks, internal links and attachments are not carried across. Page content, fonts and images are.

Is a drawn signature legally binding?

That is a question about your jurisdiction and your counterparty, not about this tool, and in many places a drawn mark on a document is exactly as binding as a scanned one. What it is not is a *digital* signature: nothing here is cryptographically bound to the file, so it does not prove the document is unaltered and it does not prove who drew it. If you need that, you need a certificate and a signing service.

Can I edit the text that is already on a page?

Yes, with one honest caveat. Pick up the text tool and every line the document holds is outlined; click one and you get a field with the original words in it, ready to change. What happens underneath is that the line is covered in the page's own background colour and the new words are written on top in a matching size, face and ink colour — because a PDF holds glyphs at co-ordinates rather than editable strings, and no browser editor can change those in place. The page genuinely shows what you typed, to a reader and to a printer. The original glyphs are still in the file for anything that extracts text, which is why this is never a redaction.

How do I use a font that is not one of the three?

Select some text and use System fonts or Font file in the Font panel. System fonts asks the browser for everything installed here — Chrome and Edge only, and it will ask your permission first. Font file opens a .ttf or .otf and works everywhere. Either way the font is written into the saved PDF, subsetted to the characters you used, so the document looks right on a machine that has never heard of it.

Why does my edited line come back in the wrong typeface?

Because the font it was set in is not loaded, so it fell back to the nearest of Helvetica, Times and Courier. Load the machine's fonts — or the specific font file — before editing, and the original's name is matched against them and the line is rewritten in its own face, in the right weight and slant.

Why does the text tool find nothing on my page?

Because there is no text on it. That page is a scan — pictures of words — and nothing can read, search, copy or edit it until it has been through OCR. The PDF OCR tool adds a text layer; after that, this tool will find the lines. It is worth knowing early, because every other thing you might try on a scan will fail for the same reason.

The cover behind my edited line is visible. Why?

Because the background there is not one flat colour — a gradient, a photograph, or a table rule running through the words — and a rectangle cannot hide that invisibly. The editor detects it and says so rather than letting you find out after sending the file. The options are to place the replacement somewhere plainer, to widen the cover until it lines up with a real edge, or to accept it.

How do I add a blank page?

Pages, then Blank page — it goes in after the page you are looking at, the same size as it. The arrow beside the button offers A4, Letter, Legal and A3, each portrait or landscape, for when the blank page is the start of fixing a document that is the wrong size. What you get is an ordinary page: write on it, stamp it, sign it, turn it, duplicate it, move it.

Does undo work per letter?

No, and that is deliberate. A run of related changes close together — typing a caption, dragging a slider, holding a key on a number field — collapses into one step, so one undo takes back the whole burst. Changing a different property, or coming back to the same one after a pause, starts a fresh step. Recording one state per keystroke would be literally correct and completely useless.

What is the difference between a sticky note and a text box?

A text box is ink: words with a panel behind them, drawn onto the page, printed with it, and visible to anything that renders the file. A sticky note is a real PDF comment — an annotation object carrying your text and your name — so readers show it as a note that opens and can be replied to, it can be turned off, and it does not print. Use a text box to change what the document says, and a note to say something about it.

What happens to a picture I place — is it re-encoded?

Once, on the way in, and then never again. A PDF can hold PNG and JPEG, so anything else the browser can open — HEIC from a phone, SVG, WebP — is converted at that point, and the bytes the mark carries from then on are exactly the bytes written into the file. Flipping and resizing are drawing instructions rather than a second conversion, so the picture never degrades from being moved around.

Why is my picture behind the box I drew?

Because the box was added after it, and objects are drawn in the order they were added. Select the picture and use Bring to front on its floating toolbar — or select the box and send it back. It is the same stack a page of overlapping objects always is; there is just nothing else to infer the order from.

Is a cover box a redaction?

No, and this matters. The box hides the words from anyone looking at the page; the words themselves are still in the file, and anything that extracts text will find them. For a real redaction, cover the text here and then run the result through Compress PDF in its redraw mode, which rasterises each page and leaves nothing but pixels.

Where are my saved signatures kept?

In this browser profile, in local storage, and nowhere else. They are not uploaded, not synced and not readable by any other site. That also means they are lost if you clear site data, and that they will not be there on a different machine or a different browser — the same fact seen from the other side.

Can I put the same stamp on every page?

Yes. Place it once, then use "put this on every page" on the object toolbar that floats above it. Each page gets its own copy, so any of them can be moved or removed afterwards without affecting the rest — and a picture used that way is embedded once however many pages carry it.

Why does my rotated mark not match its dashed outline exactly?

Only text does that, and only a little. A text mark has no box of its own — it is an anchor and some words — so the outline is estimated from the character count rather than measured with the embedded font, which is not available while you are dragging. Nothing is measured from that estimate when the file is written: the words land where the anchor says, at the angle you set.

Why did my quotation marks change shape?

The standard PDF fonts carry Latin-1 and nothing else, and curly quotes, em dashes and ellipses are outside it. Rather than refusing the save, those are folded to the nearest character the font does have. Anything with no equivalent at all — most non-Latin scripts — becomes a question mark, which is a limit of the built-in faces rather than of the file format.

Related tools

More Images & Files tools