Hey DEV community,
I’m building SciFigureAI, an AI tool that helps researchers, PhD students, educators, and science communicators create scientific figure drafts from text prompts, sketches, or existing images.
The problem I’m trying to solve is pretty specific:
Scientific figures are important, but making them is still slow.
A researcher might understand the method, mechanism, or workflow clearly, but turning that idea into a clean visual often means jumping between PowerPoint, Illustrator, Figma, screenshots, generic AI image tools, and a lot of manual cleanup.
SciFigureAI is an attempt to make that workflow faster.
Instead of being a generic image generator, the product focuses on scientific communication use cases:
- turning text prompts into scientific figure drafts
- converting rough sketches into cleaner diagrams
- editing or iterating from existing images
- creating visuals for papers, posters, slides, teaching, and lab communication
The product is still early, so I’m especially interested in feedback from people who work with research, education, scientific writing, biology, medicine, engineering, or technical communication.
A few questions I’m thinking about:
- What kinds of scientific diagrams are the hardest to create?
- Where do AI-generated visuals currently fail for serious research use?
- What export formats would actually matter: PNG, SVG, PPTX, PDF?
- Would you rather start from text, a sketch, an uploaded image, or a paper section?
You can try it here:
I’d love any feedback, especially from people who regularly create diagrams, figures, or technical visuals.
Top comments (0)