Create a clean streetwear lookbook grid showing four coordinated outfits on adult models against simple urban backdrops. Use consistent lighting, full-body poses, modern editorial styling, clear outfit variation, and a catalogue layout with no readable text.
01
Variables make prompts reusable
A fixed prompt describes one image. A prompt with variables describes a family of images. By turning details like subject, setting, product, audience, or mood into editable slots, the same structure can support many generations without rewriting the entire instruction.
For example, a prompt can reference {Subject}, {Setting}, and {Lighting}. The template page can then render those variables as fields. A user enters a value once, and the final prompt updates in a controlled way.
02
Choose variables that change often
Good variables are not random blanks. They are the parts of a prompt that naturally change from project to project. Subject, product name, scene, environment, camera angle, color direction, and intended use are common examples. The stable parts, such as composition and quality constraints, should usually remain in the base prompt.
A useful test is simple: if a user would ask 'what should I put here?', the slot needs a clear label and example. The slot called Subject might use an example like 'ceramic coffee cup' or 'founder portrait'. The example helps the user understand the level of detail expected.
03
Keep the compiled prompt visible
Variables should not hide the real prompt. The compiled prompt should be visible before generation so users can inspect, copy, or edit it. This is especially important for open templates because portability depends on clarity.
In Open Image Templates, variables sit inside an open JSON schema with slots, controls, examples, and licensing metadata. That means humans can use the template in a browser, while tools can fetch the JSON endpoint and render the same fields in another image platform.