SciFigureAI

SciFigureAI is a design tool that turns abstracts and sketches into scientific figure drafts. SciFigureAI generates scientific figure drafts, graphical abstracts, mechanism diagrams and protocol workflows from text, sketches or existing images, with editing in one workspace and PPTX or SVG export using credits.

SciFigureAI is a web-based figure maker aimed at researchers who need visuals for papers, posters, grant proposals and slides but do not want to build them from a blank canvas. A user pastes an abstract, a mechanism description or a rough visual direction, and the tool returns a clean first draft inside a project workspace. The output covers graphical abstracts, mechanism diagrams, protocol workflows, concept figures and data workflow schematics. Three input routes are offered. Text to Figure starts from a written prompt. Sketch to Figure takes an uploaded rough drawing and builds a more complete scientific visual from its structure. Edit Image accepts either a generated or an uploaded picture and revises composition and details through follow-up instructions. Because every result stays in the same project, iteration happens in context rather than through repeated one-off generations. The homepage also describes several visual directions, including schematic, 3D, photorealistic, workflow and mechanism styles. A public inspiration gallery shows example prompts across immunology, cancer biology, materials science, chemistry, physics, neuroscience, genomics, microbiology, medical imaging and ecology. Each can be copied and adapted, which helps newcomers learn how to phrase requests with details such as white backgrounds, concise labels and left-to-right layouts. The prompt box displays an estimated credit cost before generation (5 credits for the sample), so spending is visible up front. Export is handled through PPTX and SVG files, intended for slide decks and later editing in vector tools. According to the site, previews can be downloaded for free while file exports consume credits. The product positions itself as a tool for fast visual exploration followed by researcher review, so drafts are expected to be checked and refined rather than submitted untouched. Among alternatives, it sits between general image generators, which lack scientific conventions, and manual illustration suites that demand design skills and time. Labs, graduate students, clinicians and educators who produce figures regularly are the natural audience.