Dembrandt vs Lighthouse
Lighthouse audits web quality. Dembrandt audits design tokens.
Lighthouse loads a URL and scores it across performance, accessibility, SEO, and best practices. Dembrandt loads the same URL and extracts every design token the browser rendered: colors, typography, spacing, radius, shadows. Then it scores how far they drifted from the previous deploy. Same starting point, different audit.
Comparison
| Lighthouse | Dembrandt | |
|---|---|---|
| What it measures | Performance, accessibility, SEO, best practices | Design tokens: colors, typography, spacing, radius, shadows |
| Color contrast | WCAG contrast as part of accessibility score | Per-token WCAG AA and AAA grades with exact ratios |
| Token extraction | No token extraction | Full token set extracted from the live DOM |
| Drift detection | No baseline comparison or drift scoring | Numeric drift score per deploy against a pinned baseline |
| CI integration | Yes, via lighthouse-ci and GitHub Actions | Yes. --compare exits 1 on drift, so the build fails. |
| Brand regression | Not in scope | The core use case. It flags tokens that shift between releases. |
| Multi-page runs | One URL per run, results not aggregated | --crawl and --sitemap extract a whole domain in one run |
| Browser engine | Chromium only | --browser chromium or firefox, so engine differences show up |
| Agent access | None | An MCP server with fifteen tools, plus 43 design skills |
| Output | HTML report, JSON, scores 0 to 100 | JSON, W3C DTCG, CSS variables, DESIGN.md, Tailwind v4 theme, a self-contained HTML report, and a brand guide PDF |
What Lighthouse covers
- Page load performance: FCP, LCP, CLS, TBT
- Accessibility: missing alt text, low contrast elements, ARIA issues
- SEO: meta tags, canonical URLs, crawlability
- Best practices: HTTPS, console errors, image formats
What Dembrandt covers
- All computed color values with exact WCAG AA and AAA contrast ratios
- Typography: font families, sizes, weights, line heights by context
- Spacing system, border radius, box shadows
- Drift score vs baseline: what changed between this deploy and the last
Where they overlap
Both tools load a production URL and inspect what the browser renders. Lighthouse includes a contrast audit in its accessibility category: it flags elements where the text-to-background ratio fails WCAG AA and counts them toward the score.
Dembrandt extracts every color on the page and computes AA and AAA grades per token, with exact hex values and contrast ratios. You see which color fails, not how many elements do. Run both in CI. Web quality and token drift are separate gates, and each one catches what the other misses.
Run a one-off extraction on any URL with no account required:
To enforce drift in CI, get an API key, then run --compare <baseline> --key $DEMBRANDT_KEY in your deploy step. It exits 1 when the tokens move. Full setup: cloud drift CI recipe.
Other comparisons
Every tool here solves a different part of the problem. The full list says which does what.