Writing better prompts helps you get the most out of Polymet. While we provide system-level assistance, prompts written by you with real context consistently lead to better results.

Here are our best practices for crafting powerful prompts:

​
Use your product context

Instead of saying “Generate a table,” explain where the table will be used, who the user is, and what problem it solves.

Example prompt: “Create a pricing table for a SaaS product targeting startups. It should compare 3 plans and highlight features like integrations, user limits, and priority support.”

​
Describe your product, not technical details

Tech stacks in your prompt don’t help. Polymet works best when you share how your product works, who it’s for, and what kind of experience you’re building.

You can use details like:

  • What the user sees and does
  • The problem your product solves
  • Any UI/UX flows already in place
  • Preferred design styles or themes

​
Iterate part by part

Prompts that try to generate an entire product in one go often fail. It’s better to work step by step:

  1. Start with one page, e.g., the dashboard home
  2. Continue with others, e.g., analytics, settings, etc.
  3. Add interactions and flows afterward

This approach allows AI to build with better structure, clarity, and consistency.

​
Use context if needed

It’s better to include enough information inside each prompt to make it self-contained.

You can attach sentences like:

  • “I’m building an onboarding page for a fintech app. The goal is to guide new users through account setup.”
  • “We’re currently designing a user settings page where users can manage notifications and privacy.”

Even short notes like this significantly improve the quality of results.

​
Define limitations and be specific

There’s no harm in limiting the scope. The more precise the instruction, the more room AI has to be creative within those bounds.

Example: “Create a hero section with a product shot on the right and CTA on the left. No navbar yet. Targeting designers.”

​
Don’t try to generate everything in one prompt

As your project grows, avoid sending prompts that affect multiple components and pages at once.

Instead, divide large ideas into focused tasks:

  • First create the layout → Then add data sections → Then animate or style
  • Keep prompts scoped to 1 goal per prompt

This makes the model more reliable and aligned.

​
Define your style preferences early

As your project expands, changing styles becomes harder due to context limits.

Set your design tokens and visual language early: spacing, font, color, radius… so every new page aligns automatically.

​
Create flows by prompts

Once your initial designs are ready, you can define user flows by selecting an element and sending a prompt about its action.

Example: “When users click this button, open a modal with…”

​
Image prompts

If you use an image prompt, Polymet will try to recreate the visual style and layout of that image.

To go beyond that, combine the image with a descriptive text prompt for more flexibility.

​
Create multiple prototypes

You can create separate prototypes inside the same project. Each prototype can reuse the same components and pages but follow a different user flow.

By following these principles, you’ll help Polymet generate higher-quality results, reduce rework, and move faster across every stage of your design process.

If you’d like to suggest new methods that you’ve tried before, please share with us at support@polymet.ai