Scientific figures from one sentence.

promptFigure is an AI figure generator that turns one sentence or prompt into publication-ready scientific figures — mechanism diagrams, experimental flowcharts, technical roadmaps and graphical abstracts — for papers, theses and math-modeling reports.

What is promptFigure

promptFigure is an AI figure generator built for research work. Describe the subjects, relationships and layout of your figure in plain language; a server-side prompt pipeline expands that sentence into a professional drawing prompt and returns a figure you can drop straight into a paper — no drawing skill and no illustration software required.

Figures you can generate

  • Mechanism diagrams and signalling pathway schematics
  • Experimental flowcharts and protocol / workflow diagrams
  • Technical roadmaps and research framework diagrams
  • Algorithm flowcharts and model diagrams for math modeling
  • Graphical abstracts (visual abstracts)
  • Structural schematics, comparison figures and result summary diagrams

Where it fits

  • Journal paper figures (SCI / Scopus indexed)
  • Thesis, proposal and defence slide decks
  • Math-modeling contest papers and presentations
  • Lab meeting and progress reports
  • Grant applications and technical proposals

Using it for a math-modeling paper

Write the modeling steps, the meaning of each variable and the solving order as a list — what goes in, what is intermediate, what comes out — and promptFigure generates a modeling-approach flowchart, an algorithm diagram, a solving roadmap and result-comparison schematics. The hard part of a modeling paper is rarely the equations; it is communicating the approach, and that part can be handed over.

How to describe a figure

Answer three things in plain language: which subjects appear (molecules, cells, samples, models), how they relate (activates, inhibits, contains, precedes) and how the panel is laid out (left to right, top and bottom rows, hub and spoke). Attach a reference image when you have one to pin down composition and colour.

Chinese and English

Prompts can be written in Chinese or English, and labels in the generated figure follow the language you write in. The workbench, the documentation and the API are all available in both languages.

Getting started

  1. Open the workbench and describe the figure you need in one sentence
  2. Add the structures, connections and labels that must appear
  3. Optionally upload a reference image for composition and colour
  4. Verify the science in the result, then download a PNG

Frequently asked questions

What kinds of figures can promptFigure draw?

Mechanism diagrams, experimental workflows, technical roadmaps, structural schematics and graphical abstracts. Describe the subjects, relationships and layout; attach a reference image if you have one.

Do I need prompt-engineering skills?

No. Describe what you want in plain language — the server-side pipeline expands a sentence into a publication-grade drawing prompt. For complex content, list the structures and connections point by point, then verify the science in the result.

How does billing work? Plan quota vs. balance?

Plans include two pools: standard (refreshed daily, for drafts and iterations) and premium (refreshed monthly, for final renders). API usage is pay-as-you-go from your balance, separate from plan quota. See the pricing page for details.

Do subscriptions auto-renew? How do I cancel?

Monthly plans can be cancelled anytime from the console; you keep access until the end of the cycle. Single-month and yearly plans do not auto-renew. Check the terms before subscribing.

Can I download the figures?

Yes — download finished figures as PNG from the result panel, or re-download later from the gallery. Please verify text, arrows and scientific relationships before use.

Can I upload a reference image?

Yes, both the workbench and the API accept a PNG reference image (≤ 8MB) to guide composition and visual style.

Can Claude Code or another AI agent draw figures for me?

Yes. We ship a ready-made agent skill — once installed in a skill-capable AI (Claude Code / Claude Desktop / WorkBuddy, etc.), the AI handles signup, key creation and turns your description straight into API calls.

Does it work in Chinese as well as English?

Both. Write your description in Chinese or English — the workbench, the docs and the API all ship in both languages, and labels in the figure follow the language you write in.

How are mechanism diagrams and pathway figures generated?

List the molecules, structures, interactions and direction of the pathway, then describe the layout (left to right, membrane placement, arrows for activation or inhibition). The pipeline turns that into a publication-ready schematic with consistent colours and labels.

Can it draw the flowcharts and model diagrams for a math-modeling paper?

Yes. Write the modeling steps, variables and solving order as a list — what goes in, what is intermediate, what comes out — and it generates process flowcharts, algorithm diagrams, solving roadmaps and result-comparison schematics for the paper or the defence.

Can it generate a technical roadmap or an experimental workflow?

Both. A technical roadmap is best described by research stages and how the parts connect; an experimental workflow by the chronological steps — sample prep, grouping, assays, statistics. Labels can be in Chinese or English.

Open the workbench and write your first description — or install the official agent skill and let Claude Code generate figures for you.