SciFig

What is SciFig?

SciFig is an AI scientific figure generator and figure maker aimed at researchers. The site describes it as "the AI scientific illustrator for researchers": a user describes a figure in plain language, or supplies another kind of input, and receives a figure that remains editable after generation. Output is positioned for papers, presentations and teaching rather than for general-purpose image creation.

The product is presented by the SciFig Team, with contact at hello@scifig.ai. The site states it is trusted by "100K+ researchers from leading institutions," a company-reported figure that is not independently verified on the pages reviewed. The about page is marked as reviewed by the SciFig Team and updated 2026-07-23, and notes that 0.5% of purchases are contributed to Stripe Climate. No company registration or founding details are given on these pages.

How SciFig works

SciFig organizes its workflow into three stages: Generate, Enhance and Vectorize.

In the Generate stage, six input modes are advertised: Text-to-Figure, Figure Enhancer, Sketch-to-Figure, Reference-to-Figure, PDF-to-Figure and Photo-to-Figure. Generation produces output in one of six publication styles — scientific, line, 3D, editorial, sketch or watercolor — plus an Auto mode that skips the style overlay and follows the prompt alone. The site says the styles are "engineered to match Nature, Cell, and Science."

In the Enhance stage, generated figures stay editable. Users can change any text's content, font, size or position directly, open the figure in a built-in Canvas for element-level edits, supply a PDF as context for smarter edits, or upscale output to 4K/8K.

In the Vectorize stage, figures can be exported as editable PPTX or layered SVG and refined inside the built-in canvas with undo/redo, avoiding an Illustrator or Figma round-trip.

Main features

• Six input modes covering text, hand sketches, reference images, PDFs and lab or microscopy photos.
• Six publication styles plus an Auto mode.
• Every text element editable after generation (content, font, size, position).
• Built-in Canvas for editing paths, text and colors, with undo/redo.
• Vectorization to editable SVG / PPTX / Canvas.
• One-click upscaling to 4K and 8K PNG for large-format printing.
• Export formats: PNG, JPG, layered SVG and editable-shape PPTX.
• Free downloadable poster templates (A0, A1, 36x48, .pptx) and 16:9 presentation templates.
• The pricing page states data is not used to train any model and is encrypted in transit and at rest.

SciFig contrasts itself with general AI image generators: it claims scientific accuracy based on discipline-specific terminology, editable output rather than flat raster, and upscaling rather than fixed output size.

Pricing

SciFig uses credits, with monthly and yearly billing and separate credit packs. Taxes may apply.

Free — $0/month: 50 credits at signup, 50 refresh credits daily, 1GB storage, 1K HD generation. Excludes 4K/8K upscale, vector export (SVG), editable PPTX and full publication/commercial rights. Described as evaluation only; the site states publishing requires a paid plan or credit pack.

Starter — $20/month, or $14/month billed yearly ($240 vs $168 per year): 2,500 credits/month, $0.28 per figure, 10GB storage, 2K HD default, 4K upscale, SVG, PPTX, Canvas, publication and commercial rights. Modes: Text-to-Figure, Figure Enhancer, Sketch-to-Figure.

Plus — $56/month, or $20/month billed yearly ($672 vs $240): 7,000 credits/month, $0.14 per figure, 50GB storage, 4K/8K upscale, all six input modes.

Pro — $120/month, or $37.50/month billed yearly ($1,440 vs $450): 15,000 credits/month, $0.13 per figure, 300GB storage, all six input modes. The pricing page notes Plus billed annually ($240/year) costs the same as Starter monthly.

Typical credit costs: 50 credits for any generation mode or for background whitening, text editing or 4K upscale; 75 for 8K upscale; 150 per path for image vectorization, which is free afterward and unlocks unlimited free editing and free SVG/PNG/JPG/PPTX exports inside Canvas. SciFig warns that each "Edit" click on the result page (Export SVG, Export PPTX, Open in Canvas) runs a new vectorization at 150 credits, and recommends opening Canvas first and exporting from there for free.

Common use cases

• Journal figures and graphical abstracts, including vector-ready output for submission.
• Conference posters using the free A0/A1/36x48 .pptx templates, or slide decks using 16:9 templates.
• Cleaning up hand-drawn lab sketches into presentable figures.
• Converting microscopy or apparatus photos into clean diagrams.
• Turning a paper PDF into a figure brief, or matching the style of an existing publication.

Limitations

The free plan is explicitly evaluation-only: publication and commercial rights, vector export, PPTX and 4K/8K upscaling require a paid plan or credit pack. PDF-to-Figure, Photo-to-Figure and Reference-to-Figure are not available on Starter.

Costs are consumption-based. Vectorization is charged per path, and the site warns that every "Edit" click on the result page triggers a new 150-credit vectorization if Canvas is not used first. Cancellation is described as available anytime with no hidden fees, but no refund terms are stated on the reviewed pages.

On accuracy, SciFig says figures are "built on discipline-specific terminology" and that users correct anything the model gets wrong — accuracy is user-verified rather than guaranteed. No independent benchmarks or third-party validation of figure correctness are provided, and the 100K+ researcher claim is company-reported.

Summary

SciFig is a credit-based AI figure generator for researchers, differentiated by editability after generation: text-level edits, built-in Canvas, layered SVG and editable PPTX, plus 4K/8K upscaling. Pricing starts at a free evaluation tier with restricted rights, rising through Starter, Plus and Pro plans, with per-action credit costs that users should plan around — particularly vectorization.