SciFig is an online AI scientific figure generator and creation tool, mainly aimed at researchers, students, teachers, R&D teams, and science communicators, for turning complex scientific content into clear figures. According to the official description, it can convert descriptions, sketches, photos, PDFs, references, and data files into editable, ready-to-use scientific figures, including illustrations, statistical charts, and flowcharts.
Its main capabilities are divided into three engines. The first is Illustration, used for mechanism diagrams, pathways, graphical abstracts, medical illustrations, molecular structures, etc., and can be generated from a sentence, sketch, photo, or PDF. The second is DataChart, used to create statistical charts from uploaded data files; it officially supports CSV, TSV, TXT, or Excel files and claims 49 built-in publication-grade chart types, with examples including volcano plots, Kaplan–Meier survival curves, heatmaps, forest plots, raincloud plots, ROC curves, and more. The third is FlowChart, used for research flowcharts such as PRISMA and CONSORT, model architectures, cycle diagrams, and timelines, and can generate editable diagrams from a described process or pasted screening and enrollment counts.
In terms of usage, users can choose entry points such as Text-to-Figure, PDF-to-Figure, Photo-to-Figure, and Sketch-to-Figure to generate figures from scratch. After generation, the results can be placed into the Vector Canvas for editing; the official description says that clicking on a text label or shape allows direct modification, and the Vectorize feature can convert the entire figure into a layered SVG and export a PPTX with editable shapes. Common export formats include PNG and JPG; DataChart also supports PDF, EPS, and TIFF, while FlowChart supports SVG, PNG, JPG, and editable PPTX. The official statement says that 1K or 2K resolution can achieve 300 DPI at standard journal column widths, and it supports one-click upscaling to 4K/8K for large-format printing or journal covers.
Applicable scenarios include journal paper figures, graphical abstracts, systematic review flowcharts, randomized trial flowcharts, mechanism and pathway illustrations, medical illustrations, molecular structures, model architectures, timelines, and teaching slides, among others. The official statement says it supports 11 disciplines, including biology, chemistry, physics, engineering, computer science, energy, ecology, bioengineering, astronomy, agriculture, and materials, and is not limited to biology.
Regarding limitations and notes, the official description states that the free plan is for non-commercial evaluation only; paid plansprovide full publication rights and commercial use rights, with paid plans starting at $100 per year (about $8.33 per month) or $19.9 per month. As for the free quota, the official statement says there are initially up to 150 free credits, with no credit card required; joining via invitation can earn up to 350 free credits on the first day. Vectorize is a feature of paid plans and credit packs, charged at 150 credits per figure. Regarding AI hallucination, the official statement emphasizes that every element is editable, DataChart values come from user-uploaded files, FlowChart counts come from user descriptions, and the model does not fabricate numbers on its own. Regarding journal disclosure, the official recommendation is to check the target journal's policy and, if necessary, note the use of AI assistance in the methods section.
