LabFigLabFig

Vector Canvas: the scientific figure editor where your figure becomes final

Vector Canvas is the scientific figure editor that opens after a figure is generated. It is not a flat picture — the Vector Canvas is a workspace of layers, selectable text, arrows, and colors. Click any label to retype it, recolor a panel, lasso one region and regenerate it, then export the whole thing as SVG, PDF, or PNG. This is where the last 10 percent of control lives — the finishing step for everything from machine learning diagrams to physics figures and robotics diagrams.

Open the workbenchSee examples

†Generate a draft below, then refine it on the Vector Canvas

The LabFig Vector Canvas scientific figure editor with a layers panel, a selected text label, a color picker, and a regenerate-region action
An AI-generated scientific figure open on the Vector Canvas editor as live, selectable vector objects

Vector Canvas · layers, text editing, recolor, regenerate region

Layers & selection
Regenerate a regionSVG · PDF · PNG

A tour of the Vector Canvas tools

Four things you do on the Vector Canvas scientific figure editor that a one-shot image generator simply cannot. Click through to see each in place.

Every label is real, selectable text — not pixels painted onto an image. Double-click to fix a typo, rename a node, change a units suffix, or adjust the font and size. Greek letters, subscripts, and gene names stay crisp at any zoom because the text is vector, not baked in.

Before / after

Drag to see what editing on the Vector Canvas changes

On the left, a raw AI draft mid-edit — crooked labels, off colors, a misread region. On the right, the same scientific figure after a few minutes on the Vector Canvas. Same starting generation, very different result.

polished final figureAfter editing
messy mid-edit draftAI draft
Why edit on the Vector Canvas

A scientific figure editor for finishing figures, not a one-shot guess.

Every LabFig creation path — Text to Figure, Sketch to Figure, Reference, PDF, Photo, and the Enhancer — converges on the Vector Canvas. This is where a fast draft becomes the scientific figure you actually wanted.

Workspace

Layers, selection & alignment

Work like a real scientific figure editor: select any object, reorder layers, snap panels to a grid, and align labels evenly on the Vector Canvas. Direct manipulation with undo and redo — no menus you need a design degree to find.

Layers panelSelect & moveSnap & alignUndo / redo
Precision

Regenerate just one region

When a single panel is off, you do not start over. Select that region, prompt it, and the Vector Canvas redraws only that area — keeping the rest of your hard-won layout untouched.

Region inpaintingTargeted promptsLayout preserved
Output

Journal-grade export, no design skills

Export scalable SVG and PDF or a 300-dpi PNG, sized and styled to common journal standards. If you can click and type, you can ship a submission-ready figure — no Illustrator, no Figma.

SVGPDFPNG · 300dpiLayers preserved
Start a draft

Generate a first draft, then refine it on the Vector Canvas

The Vector Canvas always starts from a generation. Edit the example prompt below or write your own, hit generate, and the result opens in the scientific figure editor as editable vector objects you can relabel, recolor, and regenerate region by region.

Text → Figure
315/2000

Describe the figure you need — or add a reference image to transform.

Advanced
Cost 2 credits·Remaining 0 credits

Your figure will appear here

Pick a mode · Describe the figure · Generate

Vector Canvas FAQ

Common questions about editing scientific figures on the LabFig Vector Canvas scientific figure editor.

The Vector Canvas is LabFig's scientific figure editor. After you generate a figure — whether from text to figure, sketch to figure, a reference, a PDF, or a photo — it opens on the Vector Canvas as live, editable objects: layers, selectable text, arrows, shapes, and colors. You refine the scientific figure and export it, all in one place, without a separate design app. For the bigger picture, read our primer on what AI for scientific figures is.

Yes. The figure is genuine vector, so you can export an SVG or PDF and open it in Illustrator, Inkscape, or Affinity for further work. Most researchers finish entirely on the Vector Canvas, but nothing locks you in — the export is a standard, editable vector file.

Yes. Layers and grouping are written into the exported SVG and PDF, so panels, labels, and arrows stay as separate, selectable objects when you reopen the file elsewhere. The 300-dpi PNG is a flattened raster for when a journal specifically asks for one.

Yes — that is one of the main reasons the Vector Canvas exists. Lasso a single panel, node, or cluster and regenerate only that area with a focused prompt. The rest of your scientific figure stays exactly as you left it, and full undo and redo let you compare versions safely.

Yes. Beyond editing what was generated, you can add new text boxes, arrows, and basic shapes directly on the Vector Canvas — to annotate a panel, add a legend, or label a step the draft missed — and they export as the same editable vector objects.

Labels handle common scientific notation — Greek letters, subscripts and superscripts, and standard math symbols — and keep them crisp because the text is vector. For heavy LaTeX-style equations, lay out the figure on the Vector Canvas and refine the specific symbols before export.

Open the Vector Canvas and finish your scientific figure

Generate a draft, then use the Vector Canvas scientific figure editor to edit every label, recolor every panel, regenerate any region, and export it submission-ready — free while you explore.

Where every creation path lands · No design skills required

Vector Canvas: The Scientific Figure Editor | LabFig