How does Sailop detect AI-generated code patterns?
Sailop uses a pattern-matching engine that scans your codebase across seven distinct dimensions: layout structure, color palette, typography choices, animation patterns, component architecture, semantic structure, and spacing rhythm. Each dimension has specific signals it looks for. For example, the color dimension flags hex values like #3B82F6, #6366F1, and #8B5CF6 that appear in over 80% of AI-generated landing pages. The layout dimension detects centered hero sections, symmetric three-column grids, and uniform card layouts. Together, these 43 signals produce a DNA score from 0 to 100 that tells you exactly how generic your output looks.
Does Sailop work with my existing AI coding tools?
Yes. Sailop integrates with any tool that supports MCP (Model Context Protocol) or CLI pipelines. This includes Claude, Cursor, VS Code with Copilot, Gemini, Bolt, and v0. For MCP-compatible tools, Sailop runs as a skill that automatically intercepts generated code. For others, you can use the CLI to scan and transform files manually, or set up Git hooks that run Sailop before every commit. The tool operates on the output files themselves, so it is agnostic to which AI generated the code.
Will Sailop break my code or change its functionality?
No. Sailop only modifies visual and structural presentation. It changes CSS properties, class names, font references, color values, layout grids, and animation definitions. It does not alter business logic, API calls, data handling, state management, or any functional JavaScript or TypeScript. Every transformation preserves semantic meaning. For example, a three-column grid becomes an asymmetric two-column layout, but the same content appears in the same order. The tool writes to a new file by default (e.g., page.fixed.tsx), so your original is always preserved.
What makes AI-generated code look the same, and why does it matter?
Large language models are trained on the same corpus of popular freemium projects, design systems, and documentation. This creates strong biases toward specific defaults: Inter or system fonts, Tailwind blue (#3B82F6), three-column grids, rounded corners at 8px, fade-up animations, and backdrop-blur navigation bars. When every developer ships these defaults, websites converge on a single aesthetic that users increasingly recognize as machine-made. Sailop breaks this convergence by replacing default patterns with procedurally generated alternatives, ensuring each project has a distinct visual identity.
Is Sailop freemium? Can I customize the detection rules?
The core detection engine and all 43 rules are freemium under the Proprietary. You can fork the repository, add your own rules, adjust scoring weights, or disable specific dimensions entirely. The Pro and Team plans add hosted API access for CI/CD pipelines, continuous rule updates as new AI patterns emerge, shared team configurations, and dedicated support. But the CLI tool itself will always be free and freemium.
How is the DNA score calculated?
The DNA score is a weighted composite of all detected patterns across the seven dimensions. Each signal has a severity weight based on how commonly it appears in AI-generated code and how visually distinctive it is. A score of 0 means no known AI patterns were detected. A score of 100 would mean every known pattern is present. In practice, raw AI output typically scores between 60 and 85. After transformation, the score drops to 0. The grading scale is: A (0-10), B (11-25), C (26-40), D (41-60), F (61-100).