Start here
Create Raster Category Icons
Use one style reference to create and inspect a matching set of category icons.
StartStart here
Use one style reference to create and inspect a matching set of category icons.
StartA book, a pair of headphones, and a bookmark can each look good while looking as though they belong to different apps. When you need a set, the shared appearance and the meaning of each symbol both matter.
In this Course, you will use a reading app's prepared briefs to ask AI for matching icons, inspect the actual results, and decide what to keep or change. You will work with larger illustration icons for category cards, small editable navigation icons, and new symbols that must fit an existing set. The Lessons are self-contained: follow the suggested order or choose the task you need.
You will practise naming each icon's meaning, giving the whole set common rules, and checking the files at their intended display size. Downloads include the actual example files and reusable request worksheets. Text labels remain beside the icons; a visually consistent set does not prove that every user will understand it without words.
No. You will inspect prepared results and assemble short parts of requests. The SVG Lessons explain the few file properties you need to check and provide editable examples; you do not need to write the drawing instructions yourself.
No. The first Lesson produces pixel images. The navigation Lessons produce actual SVG shapes that can be edited as vectors. Renaming a PNG or putting it inside an SVG file does not make its outlines editable.
The guided practice uses recorded outputs and supplied materials, with no upload or live generation. To repeat a task with your own labels, use an image-generation tool for raster icons or an AI tool that can return SVG code or files for vector icons. Tool access and file handling depend on the tool you choose.
You will have practised requesting, checking, and extending a set. The example files can be saved, but your own interface still needs a check in its actual layout and with its users. The Course does not measure recognition rates or interface performance.