Loading page
About this tool
Fill-in-the-blank AI prompts for moodboards, critique, UX copy and design system decisions, ready to copy.
The Designer Prompt Pack is a library of 12 fill-in-the-blank AI prompts covering the design tasks that most often produce vague output: moodboard direction, colour palettes, landing page structure, empty states, microcopy, critique, accessibility review, tokens, component specs and research scripts. Each prompt already states the role, the constraints and the exact output format, so the model returns a structured answer instead of a paragraph of generalities. You fill the blanks in your browser, the tool substitutes them and shows the estimated size, then you copy the finished prompt into whichever assistant you use.
Open Designer Prompt Pack on AltFTool — it loads instantly in your browser.
Add your input to the workspace.
Adjust the options until the result looks right.
Copy or download the output and put it to work.
Every prompt fixes the role, the output format and the rules, which is where most improvised prompts fall apart.
Blanks you leave empty stay visible as {{markers}} and are listed above the result, so nothing ships half-written.
Prompt assembly happens in the browser; no account, no API key and nothing you type leaves the page.
Give it four things: a role, the specific artefact, the constraints it must respect, and the exact output format you want back. A prompt that says 'critique this screen' returns taste; one that says 'critique in this order, give observation, user impact and one concrete change, and separate blocking issues from polish' returns something you can act on.
WCAG 2.2 level AA requires at least 4.5:1 for normal body text, 3:1 for large text (18pt, or 14pt bold), and 3:1 for user interface components and graphical objects. The palette prompt asks the model to report the ratio for every pair, but always re-check the numbers in a real contrast checker because language models routinely miscalculate them.
No. It is useful as a first pass that catches structural problems such as unclear hierarchy, missing states and accessibility gaps, which frees a human reviewer to spend their time on judgement, context and craft. It has no knowledge of your users, your roadmap or the political constraints around a decision.
Any current general-purpose assistant handles them, because the structure does the work rather than a model-specific trick. The tool shows an estimated token count at roughly four characters per token so you can tell when a filled prompt is getting long for a small context window; anything above about 700 tokens is worth checking before you paste it into an 8K-context model.