Rams is an automated design review tool that checks your UI code for accessibility issues, design token violations, animation problems, and AI-generated code patterns. Think of it as a linter for visual design quality.
Two ways: Use the /rams skill in your AI editor (Cursor, Windsurf, Claude) for interactive reviews while coding. Or add our GitHub Action to automatically review every pull request and block PRs that don't meet your design standards.
80+ rules across 8 categories: accessibility (WCAG compliance, focus states, alt text), color tokens, animation/motion preferences, spacing/layout, typography, component patterns, UX states, and anti-AI slop detection.
The /rams skill is completely free — it's just a static file you download. The CI/CD GitHub Action is free during beta. We'll announce pricing before general availability.
ESLint checks JavaScript syntax and code quality. Rams checks visual design quality — things like color contrast, missing focus states, hardcoded values that should be tokens, and patterns that make your UI feel "AI generated."
Yes. For the /rams skill, everything runs locally in your AI tool — we never see your code. For CI/CD, code is analyzed in memory and immediately discarded. We never store your source code, only usage metrics (file count, duration, score).
For the /rams skill: nothing — it's a static text file with no network requests. For CI/CD: UI files are sent via HTTPS to our Worker API, analyzed by Claude, then immediately discarded. Results are posted as a PR comment.
We use Anthropic Claude for analysis. Claude is SOC 2 Type II certified and does not use your code for training. All communication is encrypted with TLS.
Minimal permissions: read repository contents and write PR comments. We don't need access to your organization settings, secrets, or other repositories.
Yes — contact us about self-hosted deployment options for enterprise customers with strict compliance requirements. All infrastructure providers are SOC 2 certified.
Rams works with React, Vue, Angular, Svelte, Next.js, and vanilla HTML/CSS. We analyze JSX, TSX, Vue SFC, and standard CSS/SCSS files.