How FigureLabs Turns Inputs into Publication-Ready Figures

Updated

FigureLabs is an AI scientific figure generator that converts text, sketches, lab photos, Excel files, PDFs, or reference figures into publication-ready illustrations, flowcharts, and plots. The workflow has four stages — Input, Generate, Edit, Export — and the output can be exported as editable PPTX or SVG, high-resolution PNG or JPG, or shared via link. It suits researchers who need journal- or grant-ready visuals but want to keep them editable rather than locked in a flat image.

What you can feed in

The Input stage accepts a wide range of source material, so you don't need to start from a clean description:

  • Text — a written description of the figure you want
  • Sketches — rough hand-drawn layouts
  • Lab photos — images of experimental setups or results
  • Excel — spreadsheet data for plots
  • PDFs — documents containing figures or data to work from
  • Reference figures — existing figures whose style or structure you want to follow

According to FigureLabs, the system "understands them all," meaning you can mix formats depending on what you already have on hand.

Generating the figure

At the Generate stage you choose how the figure gets produced:

  1. Pick an AI agent — FigureLabs supports agent-based workflows, including an integration path for ChatGPT and Codex (the site points to a GitHub AGENTS.md file for installing the Codex plugin and running its default connection prompt).
  2. Pick a model or visual style — the style selection shapes the look of the output.
  3. Choose the figure type — Illustration, Flowchart, or Plot, each with its own starting point.

The three figure types map to common research needs:

Type Typical use
Illustration Graphical abstracts, signaling pathways
Flowchart Methods diagrams, PRISMA-style flows, grant proposals
Plot Bar plots and other data visualizations from raw data or statistical results

Editing on the infinite canvas

Once generated, figures land on an infinite canvas where you refine them. Available edits include:

  • Redraw selected areas — fix a specific region without regenerating everything
  • Edit text — correct labels and annotations directly
  • Recolor — adjust colors to match journal or institutional style
  • Remove backgrounds — clean up images for compositing
  • Upscale — increase resolution for print

For plots specifically, you can "refine every detail with AI chat," so adjustments can be described in natural language rather than done manually.

Two supporting features matter if you're producing multiple figures:

  • Visual Consistency — lock icons, arrows, fonts, and style so every figure in a paper matches.
  • Work Together — invite collaborators to comment and suggest on the canvas in real time.

Exporting and sharing

The Export stage gives you format choices depending on downstream use:

  • PPTX or SVG — fully editable, useful when a journal, co-author, or presenter needs to modify elements
  • PNG or JPG — high-resolution raster output for submission or slides
  • Share link — for collaboration rather than file transfer

FigureLabs also offers an Authorization Certificate: a verifiable PDF with a unique ID, which can serve as proof of creation.

A practical walkthrough

Suppose you need a graphical abstract for a signaling pathway paper:

  1. Upload a rough sketch of the pathway plus a short text description.
  2. Select Illustration as the figure type and choose a visual style.
  3. Generate, then on the canvas redraw the arrow connections, edit the protein labels, and recolor to match your journal's palette.
  4. Lock the icon and font style so your other figures stay consistent.
  5. Export as SVG for your co-author and PNG for submission.

The same four-stage flow applies whether you're building a PRISMA flowchart from a PDF or a bar plot from an Excel sheet — only the input and figure type change.

What to check before relying on it

  • Pricing and account requirements aren't specified in the available page content, so confirm current terms on the pricing page before committing.
  • Agent integrations (ChatGPT, Codex) involve setup steps outside the main app — the Codex path requires following the linked GitHub instructions.
  • Editability varies by export format: PPTX and SVG preserve editable elements; PNG and JPG do not.

If your goal is a figure you can hand to a co-author or journal in an editable format, the PPTX/SVG export path is the one that matters most. If you only need a final image, PNG/JPG is sufficient.