Handwriting from typed text

At a glance, on a printed or photographed sheet, usually not — three things work towards that: every letter has its own shape, the lines stand on the ruling, and the stroke thickness and ink density wander. Under a careful comparison the differences remain, and they are worth knowing in advance. Joins between letters come from the typeface rather than being drawn by a pen, so in connected hands the junctions are tidier than a person’s. We simulate the hand tiring across a page, but not mood, haste or corrections. And the decisive one: it is not YOUR hand — next to your own signature or another note of yours the difference shows immediately. If the result has to match your own writing, capture it under “My handwriting”. The tool is not meant for documents where the authenticity of the hand carries legal weight.

Because not every handwriting typeface covers Cyrillic, and offering one that would draw empty boxes instead of letters is worse than not offering it. The list is rebuilt for the writing system your text actually uses: type in Russian and only the hands that fully cover Cyrillic remain; clear it and type in English and the list grows. If your text contains a script that none of the hands covers, the page says so plainly instead of leaving you to puzzle over rectangles. Korean and Arabic work the same way — there is one hand for each, and it appears only when it is needed.

It keeps the ink only and leaves the paper transparent, and it is the one format you can lay over somebody else’s document while still seeing the original form underneath the writing. An ordinary PNG or a JPEG brings a white rectangle with it and hides everything below. From there you insert it as a picture into Word, a PDF editor or a slide and move it where it belongs. If the form is a PDF, the “Fill a PDF” tab does the same job more cleanly: there the writing enters the document as vectors and does not depend on the resolution of an image. Note that “Look scanned” is not applied to the transparent variant — its edge shading would paint a grey rectangle exactly where the transparency is needed.

They answer different questions, and confusing them wastes your time. The Russian cursive page answers “how is this written”: it takes one word, animates the stroke order, gives an alphabet table with audio and a printable tracing worksheet — a school tool for someone learning to write. This one goes the other way: it takes a finished document and turns it into a page. The input is paragraphs rather than a word; the output is a sheet rather than a letter. If you want to know how a letter is formed, go there; if you want notes or a filled form, stay here.

Because there is no text in it — there are lines. Every letter is stored as a set of curves, and it cannot be otherwise: the whole point of the tool is that the same letter has a different shape each time, while text in a PDF is a reference to one glyph of a font, that is, exactly one shape for every occurrence. The consequences follow: you cannot search such a PDF and you cannot copy text out of it, but you can zoom as far as you like and the lines stay sharp. The “Fill a PDF” tab behaves differently: your original document is left untouched, so its own printed text remains text, searchable and copyable, and only what you wrote is added as a handwritten layer.

Almost always the photo rather than the handwriting. The letters are read from their outline, which needs even contrast: write with a dark pen rather than a pencil or a pale refill, shoot in diffuse light, without the shadow of your own hand and without a lamp reflecting off the paper. The whole sheet must be in frame — the four black corner squares are what the perspective correction uses, and if even one is cut off the reader refuses rather than quietly skewing every letter. Letters that ran outside their box are cropped at its edge. If the photo still will not cooperate, write the difficult letters with a finger in the other lane: there the full path of the pen is recorded and the result is cleaner than any photograph.