Figran — AI Scientific Figure Generator for Research Papers

What is Figran?

Figran is a web-based AI tool that turns research descriptions into scientific illustrations. It is aimed at biomedical and laboratory researchers, students, and educators who need figures for papers, posters, and teaching material — signaling-pathway diagrams, method flowcharts, data plots, and graphical abstracts — without drawing them element by element in Illustrator or PowerPoint.

The core interaction is textual: you describe the objects, their relationships, and the labels, and the system renders a figure. You then refine the result with follow-up instructions ("make the arrow dashed," "relabel the nucleus"), and each edit creates a new version while preserving the original, so iterating does not destroy earlier drafts. The site also supports reference-image uploads, sketch-to-figure conversion, and an editable canvas for flowcharts and data plots, including importing CSV data to produce plots.

How Figran Works

The workflow has three steps. First, describe your figure — the objects, their relationships, and any labels. Second, review the generated result and refine it via follow-up text instructions; every revision is versioned. Third, download the figure in the format you need: PNG (without watermarks on paid tiers), SVG for vector workflows, or an editable PPTX produced from the canvas tools.

Generation runs on a choice of models, including GPT Image 2 and 2.5 variants and Nano Banana Pro and 2, with Seedream 4.5 and 5.0 variants reserved for paid tiers. Credit costs vary by model and resolution — for example, GPT Image 2 costs 10, 15, or 25 credits for 1K, 2K, or 4K output — and credits reserved for failed generations are released back. Generation history is kept for 30 days.

Main Features

• Text-to-figure generation — describe objects, relationships, and labels in natural language and receive a scientific illustration.
• Versioned text-guided editing — refine figures with follow-up instructions; each edit saves a new version and preserves the original.
• Sketch-to-figure and reference images — convert a hand sketch or steer style with uploaded references.
• Editable canvas — flowchart mode and plot mode, including CSV-to-plot, for diagrams that stay editable rather than baked images.
• Publication-ready exports — PNG up to 4K, watermark-free on paid tiers, plus vector SVG and editable PPTX export.
• Multiple models and 30-day history — several image models to choose from, with a month of generation history for revisiting drafts.

Pricing

Figran uses a freemium credit model. Eligible new accounts receive 20 welcome credits valid for 30 days, plus 10 credits per 24-hour cycle (starting 24 hours after activation, subject to availability). Paid plans advertise a 50% saving on annual billing: Pro costs USD 10/month (USD 120 billed annually) with 1,200 credits per month — up to about 120 images — and 5 models; Ultimate costs USD 20/month (USD 240 annually) with 3,000 credits (about 300 images) and 8 models; Max costs USD 40/month (USD 480 annually) with 7,000 credits (about 700 images) and 8 models. Credit Packs are also available for one-off needs; pack credits last 180 days, while subscription credits do not roll over between months.

Common Use Cases

• Signaling-pathway and mechanism diagrams — the flagship use case for biomedical papers and grant applications.
• Graphical abstracts — a single polished figure summarizing a study for journal submission.
• Methods and workflow flowcharts — step-by-step experimental procedures drawn on the editable canvas.
• Data plots from CSV — turning spreadsheet results into presentation-ready charts.
• Teaching material — labeled diagrams for lectures and lab training.

Limitations

Figran's AI images remain raster — they are not automatically split into editable shapes or text, so deep editing requires the canvas tools or external editors. The site is explicit that the tool does not validate experiments or replace scientific review, and users must check their target journals' rules on AI-generated figures. Higher resolutions and premium models require a subscription or Credit Pack balance. Subscription credits do not roll over, so unused monthly credits are lost. Finally, the operating company, its country, and a launch date are not disclosed on the site — only a 2026 copyright notice and a support email.

Summary

Figran is a focused tool for a well-defined audience: researchers who need decent scientific figures quickly and can describe them in words. Its versioned text-guided editing, CSV-driven plots, and SVG/PPTX exports fit real academic workflows better than generic image generators, and the free daily credits lower the barrier to trying it. The honest caveats — raster output, no scientific validation, non-rolling subscription credits, and an undisclosed operator — make it best treated as a drafting accelerator whose outputs still need expert review before publication.