How to Put a Vizcept Figure into Word, PowerPoint or LaTeX

Every figure can be downloaded as a PNG (2K, or 4K on a paid plan) or as an SVG. Pick by where the figure is going.

Two things to know before you plan around the SVG:

• Syntax trees, sentence diagrams and conceptual frameworks are normally typeset by code, so their SVG keeps real text — you can select a label and restyle it. How Vizcept draws figures explains when they fall back to the image model instead.
• The other figure types are painted by an image model, and their SVG is traced from the finished picture. It scales cleanly, but the words in it are outlines, not text: you cannot retype a label in Illustrator, only redraw it. Fix wording by editing the figure and generating again, before you download.

Free downloads carry a visible watermark. The full-size file without a watermark, and the SVG, come with a paid plan or an unexpired credit pack.

Microsoft 365 and Word 2019 or later, on Windows and Mac, place SVG files directly:

1. Insert → Pictures → This Device, then choose the .svg file.
2. Resize by dragging a corner handle, or open the Size dialog and keep Lock aspect ratio ticked. Dragging a side handle stretches the labels.
3. To recolour or delete a part, select the figure and use Graphics Format → Convert to Shape. This is worth doing on a code-drawn syntax tree; on a traced SVG it explodes into thousands of shapes and can make the document very slow — use the PNG instead.
4. Add the caption with References → Insert Caption, not inside the image. Keeping the caption as Word text means it renumbers itself and the journal can reflow it.

If your Word version shows a blank box, or you are in Google Docs, use the PNG. Nothing is lost: at the size a figure occupies in a manuscript, a 2K PNG is normally indistinguishable from the vector.

What matters is not the number in the download menu but the pixels spread over the printed width:

effective resolution (ppi) = pixel width of the file ÷ printed width in inches

A 2K download is roughly 2,000 pixels on its long edge and a 4K download roughly twice that, but the exact size depends on the aspect ratio you picked — so check the file rather than assume. On Windows: right-click → Properties → Details. On a Mac: select the file and press ⌘I, or open it in Preview and choose Tools → Show Inspector.

If the number falls short, download again at 4K, or place the SVG, which has no fixed pixel size. Publishers usually ask for at least 300 ppi for photographs and considerably more for line art; your target journal's artwork guide gives the number it enforces.

One setting quietly undoes all of this: Word compresses images when it saves. Turn it off once, in File → Options → Advanced → Image Size and Quality: tick Do not compress images in file and set the default resolution to High fidelity.

\includegraphics reads PDF, PNG and JPG — not SVG. Three ways round it, best first:

1. Convert the SVG to PDF, then include the PDF. Works everywhere, including Overleaf.

inkscape figure.svg --export-type=pdf --export-filename=figure.pdf

(Inkscape 1.x; or open the SVG in any vector editor and Save As PDF.) Then:

\includegraphics[width=\linewidth]{figure.pdf}

2. The svg package. \usepackage{svg} and \includesvg[width=\linewidth]{figure} convert the file at compile time — but that needs Inkscape installed and the compiler run with -shell-escape. Overleaf does not run Inkscape for you, so this fails there.

3. Just use the PNG. \includegraphics[width=\linewidth]{figure.png} is fine for a draft; before you submit, check the effective resolution above.

Vizcept does not export LaTeX code. There is no forest, qtree or TikZ output, including for syntax trees — if your course requires tree code rather than an image, you will have to typeset it yourself.

Vizcept exports PNG and SVG. The two formats journals name in their production guidelines are one conversion away:

Converting does not add detail: a 300 ppi tag on a small PNG does not make it a 300 ppi figure, it only labels it. Start from the 4K PNG or from the SVG.

Many journals accept PDF or PNG at submission and only ask for production files once the paper is accepted, so check the author guidelines before you spend time converting.

The caption belongs in your document, not in the image, and if the figure was drawn with AI most publishers want that said somewhere.

• Which publisher wants what, quoted from their own policy pages, plus a generator for the caption line and the AI declaration: AI-generated figures in journals.
• Who owns the figure and what you are allowed to publish: who owns figures made with Vizcept.
• Still revising? Turning supervisor comments into a revised figure covers the rounds before this one.