Is the AI design assistant free?

Zeplik is free to start and every new account includes credits. Design skills run like any other Zeplik chat, so you spend only normal usage credits, with the cost shown before you send.

Can it critique a screenshot or a Figma design?

Yes. Paste a screenshot, the markup, or a Figma spec and it critiques what it sees: hierarchy, spacing, contrast, and affordances, each with a specific fix. It reads what you paste rather than opening your live file.

Can it really audit for accessibility?

It runs a structured review against WCAG 2.2 success criteria and reports violations by level with remediations. It is honest that an automated audit still needs testing with screen readers and real users for true conformance.

What do I need to give it?

For a critique, a screenshot or the markup. For a component, describe the UI and platform. For a system audit, paste the components or tokens. For brand work, share your colors and type.

Can it write the frontend code too?

Yes. It can turn a design or Figma spec into working frontend code, and produce component specs with all their states. Review and test the output in your own stack before you ship.

Will the design just look generic?

Not if you give it direction and context. It designs with intentional aesthetic choices and works from your brand, and you steer the fidelity and constraints. Treat the result as a strong first pass to make your own.

AI design assistant

Paste a screenshot, the markup, or a Figma spec and get design work you can act on: a critique with the specific fixes to make, a component and its states, a design-system audit, an accessibility report against WCAG, UX copy that reads clearly. Zeplik works from the screens and brand you give it, not a generic template, and it is honest about the line: it critiques and drafts, and it tells you what to validate with real users and an accessibility scan before you ship.

See pricing

Who it is for

Designers who want a fast second opinion and a structured critique
Frontend engineers building UI who need patterns, states, and accessibility right
Founders and solo builders designing a product without a design team
Teams standardizing on a design system and tightening accessibility

How it works

Share the screen or the brand
Paste a screenshot, the component markup, a Figma spec, or your brand colors and type. Nothing is assumed. The work is built only from what you share, and the more real context you give, the more specific and less generic the feedback.
Ask for the design work you need
Say it plainly, for example "critique this screen" or "audit this page for accessibility." Zeplik routes to the right design skill without you naming it, and you can steer the platform, the fidelity, and the constraints.
Review, validate, and ship
You get a ready-to-use critique, component, or audit with the reasoning shown. Validate the calls with real users and an accessibility scan, then hand it off, all in one conversation.

What you can ask

Real asks in plain language, and the concrete deliverable you get back. Open the assistant to try your own; nothing runs until you send it.

Critique this screen and tell me what to fix before I ship it.

A structured critique of the screenshot or markup you paste: hierarchy, spacing, contrast, and affordance issues, each with a specific fix and why it matters, ordered by impact.

Audit this page for accessibility and WCAG issues.

A WCAG conformance report over what you share: violations by success criterion and level, the users each one blocks, and concrete remediations you can hand to engineering.

Design a component and all of its states for the UI I will describe.

A component spec with the default, hover, focus, active, disabled, loading, and error states, plus interaction and accessibility notes, ready to build.

Audit our design system for inconsistencies.

A design-system review that finds naming drift, hardcoded values, and off-token colors and spacing across the components you paste, with a path to consolidate.

Rewrite the UX copy on this flow to be clearer and on-brand.

Revised microcopy for the buttons, empty states, errors, and confirmations in the flow you share, tighter and consistent, with the reasoning for each change.

Match the visual style of a site I like onto my project.

A breakdown of the reference site's color, type, and spacing language and how to apply it coherently to your own UI, without copying its layout wholesale.

What it can do

The assistant routes to the right skill for you. Each one below is a focused, ready-to-run workflow; follow a link to see exactly what it does.

Design interfaces

Build distinctive, usable UI with intentional direction.

UI/UX quality systemApply a full quality system to app interfaces: styles, palettes, heuristics, audits.
Frontend designDesign a distinctive app UI or component with intentional aesthetic direction.
UI patternsBuild components, state, and CSS systems the right way.
Responsive layoutsContainer queries, fluid type, grid, and mobile-first breakpoints.
Mobile designiOS and Android decisions: touch targets, platform patterns, offline UX.
Interaction designMicro-interactions, motion, transitions, and loading and hover states.
Premium web designHigh-end marketing and landing pages that must look expensive.

Critique and design systems

Get a structured second opinion and keep the system consistent.

Design critiqueStructured feedback on a mockup, screenshot, or Figma link.
Style matchingReplicate a reference site's visual style onto your project.
System auditAudit, document, or extend a design system and find inconsistencies.
Build a systemComponent libraries, design tokens, theming, and starter scaffolds.
Theme factoryGenerate a coordinated color and font theme for an artifact.
Static designPosters, covers, prints, and flyers delivered as a png or pdf.

Accessibility and UX copy

Make it usable for everyone and clear in every state.

Accessibility reviewA quick a11y check of a design or page for common issues.
WCAG auditA formal WCAG 2.2 conformance audit with levels and remediation.
UX copyWrite or review microcopy: labels, errors, empty states, and CTAs.
User researchPlan, run, or synthesize UX research, interviews, and surveys.

Brand, Figma, and handoff

Ground the work in your brand and get it to engineering.

Personas and journeysData-backed personas, journey maps, and usability-test plans.
Brand guidelinesApply a consistent brand look and feel to an artifact.
Brand landing pageA brand interview, then deployment-ready landing-page HTML.
Figma specsInterpret and organize pasted Figma specs, tokens, and exports.
Figma to codeTurn an attached design or Figma spec into working frontend code.
Developer handoffA handoff spec: layout, tokens, component props, and interactions.
Delivery workflowChain research, design, systemize, accessibility, and handoff.

What a general chatbot gets wrong on design, and what is different here

A raw chatbot will happily praise a design and call a page accessible without checking anything. Zeplik works from the actual screen you paste, applies real heuristics and WCAG criteria, and is honest that a critique is not a usability test.

The failure. A general model gives vague, agreeable feedback ("looks clean, maybe more spacing") with nothing specific to act on.

How Zeplik grounds it. Zeplik critiques against real heuristics and returns specific, prioritized fixes tied to what it sees in your screen, with the reason each one matters, so you know exactly what to change.

The failure. Asked if a page is accessible, a general model says yes without evaluating contrast, focus order, or any WCAG criterion.

How Zeplik grounds it. The accessibility skills evaluate against WCAG 2.2 success criteria and report violations by level with remediations, and they are clear that an automated review still needs testing with assistive tech and real users.

The failure. A general model invents design-token names or component states that do not match your system.

How Zeplik grounds it. It works from the components and tokens you paste, flags naming drift and hardcoded values it actually finds, and does not fabricate a system you did not share.

What it does not do

Knowing the boundary is part of using it responsibly. This is where a person, not the assistant, has to own the call.

A critique is not a usability test. Structured feedback and heuristics catch a lot, but they do not replace watching real users. Validate the important calls with actual research before you commit.
Accessibility still needs real testing. It evaluates against WCAG criteria from what you share, but true conformance needs testing with screen readers, keyboards, and assistive tech on the live product.
It reads what you paste, not your file. It works from a screenshot, markup, or spec. It does not open your live Figma file or design tool, so bring the screen or the tokens you want it to use.
Taste is a starting point. Aesthetic direction is a draft to react to, not a mandate. Treat it as a strong first pass and make it yours.

This is design critique and draft work for your review, not a finished, tested product. Accessibility, brand, and usability judgments are only as good as the screens and context you provide, so validate them with real users and an accessibility audit, and confirm anything you ship meets your brand and WCAG requirements.

Frequently asked questions

Is the AI design assistant free?: Zeplik is free to start and every new account includes credits. Design skills run like any other Zeplik chat, so you spend only normal usage credits, with the cost shown before you send.
Can it critique a screenshot or a Figma design?: Yes. Paste a screenshot, the markup, or a Figma spec and it critiques what it sees: hierarchy, spacing, contrast, and affordances, each with a specific fix. It reads what you paste rather than opening your live file.
Can it really audit for accessibility?: It runs a structured review against WCAG 2.2 success criteria and reports violations by level with remediations. It is honest that an automated audit still needs testing with screen readers and real users for true conformance.
What do I need to give it?: For a critique, a screenshot or the markup. For a component, describe the UI and platform. For a system audit, paste the components or tokens. For brand work, share your colors and type.
Can it write the frontend code too?: Yes. It can turn a design or Figma spec into working frontend code, and produce component specs with all their states. Review and test the output in your own stack before you ship.
Will the design just look generic?: Not if you give it direction and context. It designs with intentional aesthetic choices and works from your brand, and you steer the fidelity and constraints. Treat the result as a strong first pass to make your own.

More on Zeplik

Critique your next screen

Paste a screen, a component, or your brand and get a prioritized critique, an accessibility audit, or a component you can build. Free to start, and honest about what to validate with real users before you ship.

Paste a screen or describe the UI Browse all tools