GLMImage

GLMImage generates labeled scientific illustration drafts from text prompts across nine disciplines, from botany to astronomy. It targets educators, writers and researchers who need quick figure drafts for expert review. New accounts get 2 free images.

GLMImage's scientific illustration generator turns a plain-language prompt into a labeled figure draft. A user describes a specimen, structure or process, and the tool returns an educational layout with readable labels, directional arrows and, where the prompt calls for it, a cross-section, cutaway or life reconstruction. The page positions the output as an accuracy-first draft meant for reference checking and expert review, not as a finished, verified figure. The workflow is organized around nine discipline recipes: botany and plant anatomy, medical anatomy and physiology, cell and molecular biology, engineering and technical systems, astronomy and astrophysics, zoology, natural history and reconstruction, geology and earth systems, and ecology. Each recipe carries its own conventions. Medical figures use clinical tissue rendering with leader-line labels, molecular figures use flat vector pathways with arrows that separate reaction from transport, and engineering figures show connected components with flow direction. Labels can be rendered in English or Chinese, and layouts lean landscape, which suits articles and lessons. In practice, a visitor picks a recipe or opens the generator, types a prompt of up to 2000 characters, chooses a style and aspect ratio, and signs in to generate. A gallery of shared examples, such as a heart cutaway, a DNA replication fork and a wind turbine nacelle, shows the full prompt behind each image and offers a Create Similar option, which makes it easy to adapt a working prompt. The examples reveal that prompts are wrapped in long direction text that favors a clean white field, a short title, a handful of essential labels and a safe margin around the artwork. GLMImage describes itself as an independent product that may route requests to supported production models, with the model recorded on each task treated as authoritative. It sits between general image generators, which tend to garble labels, and manual tools such as vector editors or dedicated scientific illustration software. It suits teachers, students, science writers and researchers who need a fast first draft to refine, check or redraw, and who accept that the output still requires verification by a subject expert.