Art to Web Analysis

Produce specifications so precise any developer—without seeing original—can implement pixel-perfect web replica.

Core Principles

1. Framework-agnostic: CSS property names, semantic HTML. No Tailwind, React, Vue syntax.
2. Pixel-perfect: Exact measurements, colors, ratios. Uncertain → ranges + estimate.
3. Only visible: Document what's seen. No assumptions about responsive/hidden states.
4. Top-to-bottom, outer-to-inner: Visual reading order, then nested drill-down.
5. Blind assumption: Reader has zero visual access.

Technical Terminology

Use freely:

• CSS properties: letter-spacing, font-weight, line-height, gap, padding, margin
• CSS values: 0.1em, 400, #XXXXXX, rgba(), translateX()
• CSS layout: flexbox, grid, absolute, fixed, sticky
• HTML tags: <sup>, <sub>, <nav>, <button>, <input>
• CSS functions: cubic-bezier(), linear-gradient()

Avoid:

• Tailwind: tracking-wide, flex-row, gap-4, text-lg
• React/Vue/Svelte: <Button />, v-if, {#each}
• Framework patterns: "use a component", "add a hook"

Clarification

Use AskQuestion tool to resolve ambiguity:

• Section boundaries unclear? Ask user to confirm logical divisions
• Component reuse uncertain? Ask if similar elements are same component
• Interactive behavior ambiguous? Ask about intended states/triggers
• Purpose unclear? Ask what the section/element is for

Section Structure

Each section includes:

• Container: edge behavior, dimensions, padding, background
• Layout: model, direction, alignment, gap
• Elements: tag, position, dimensions, content, typography, colors, border, shadow
• Icons (if present): shape description, size, color, stroke width
• Animations (video only): target, trigger, timing, keyframes, stagger, transitions

For detailed templates, see templates.md.

Measurement Guidelines

• Colors: exact hex
• Sizes: nearest 4px
• Uncertainty: "~XXpx (estimated)"
• Common ratios: 16:9, 4:3, 1:1, 3:2

What to Document

• Every visible element
• Exact text content (character-by-character accuracy)
• All colors, spacing, typography
• Visual hierarchy, nesting
• Icons (describe visual form)
• States only if shown

What to Skip

• Image content (use category-tagged placeholders: "[Product photo]", "[Hero banner]", "[Avatar]")
• Assumed responsive behavior
• Assumed hover/focus states (only document if visible)
• Implementation details
• Accessibility assumptions

Typography Edge Cases

Document explicitly when present:

• Superscripts/subscripts: M<sup>c</sup>, 1<sup>st</sup>, TM, ®
• Small caps: font-variant: small-caps
• Mixed case patterns: note exact text-transform value
• Ligatures: font-variant-ligatures value if non-default

Text Accuracy

Critical for implementation fidelity:

• Copy exact strings character-by-character
• Preserve line breaks as visually rendered
• Include all punctuation and symbols (+, >, <, ©, •, |)
• Note year values exactly as shown
• Include country/region qualifiers
• Document trailing/leading whitespace if significant

Icon Identification

Platform/brand icons: Name explicitly (Instagram, TikTok, Pinterest, Facebook, YouTube, X/Twitter, LinkedIn)

• Style: filled, outlined, monochrome, circled, bare, branded colors
• Order: Document left-to-right sequence
• If uncertain: Describe distinguishing shape

Generic icons: Describe geometric form precisely, note stroke vs fill style.

Content Hierarchy

For grouped content (navigation, footer columns, lists):

• Headings: Document separately from items, note if styled differently
• Items: List in visual order
• Format: "Heading: [text]" → "Items: [A], [B], [C]"
• Nesting: Indicate parent-child relationships clearly

Interaction States

If multiple states visible, document each:

For each: document which properties change + transition timing if animated.

Video Analysis

• Watch multiple times at normal + slow speed
• Note every motion, transition, state change
• Capture timing precisely (video timestamps)
• Document trigger conditions
• Describe easing (snappy, smooth, bouncy)

Quality Checklist

Structure

• Every visible element documented
• Sections top-to-bottom
• Elements outer-to-inner
• Column/section headings documented separately from items

Typography & Text

• All text strings verified character-by-character
• Typography fully specified (family, size, weight, line-height, letter-spacing, transform)
• Special typography noted (superscripts, small caps, mixed case)
• Year values, country codes, symbols exact

Visual Properties

• All colors have exact hex
• Spacing for all gaps/padding
• Images as category-tagged placeholders

Icons

• Platform/brand icons identified by name
• Icon style documented (filled, outlined, circled)
• Icon order documented (left-to-right)

Animations (video only)

• Animations with timing/keyframes
• Playback controls documented if visible

Interaction States (if visible)

• All visible states documented
• State transitions noted with timing

Standards

• Uses CSS property names, avoids framework-specific syntax
• No unseen state assumptions

References

CSS property lookup:

• CSS Reference: https://developer.mozilla.org/en-US/docs/Web/CSS/Reference
• Font features: https://developer.mozilla.org/en-US/docs/Web/CSS/font-feature-settings
• Transitions/Animations: https://developer.mozilla.org/en-US/docs/Web/CSS/CSS_transitions