Show HN: StyleSeed – a design-rules engine so AI agents stop building generic UI (github.com)

StyleSeed has been unveiled as a groundbreaking design-rules engine aimed at enhancing the UI quality generated by AI agents, such as Claude Code and Codex. By integrating a system of 74 visual design rules and a scored Quality Gate, StyleSeed combats the common "generic AI look" that arises from standard model outputs. It enables AI systems to deliver coherent, brand-specific UIs while ensuring consistency across design iterations — critical for developers and businesses who prioritize professional aesthetics in AI-generated applications. The tool allows designers and developers to set specific design parameters (like key colors, fonts, and motion) that persist across session prompts, effectively preventing 'design drift.' With features such as a quality scoring mechanism (requiring a score of 80/100 for displayed outputs) and easy installation via a simple command, StyleSeed positions itself as an essential resource for anyone leveraging AI in design workflows. By emphasizing design judgment rather than mere data collection, StyleSeed elevates the capabilities of AI agents, making it a significant advancement for the AI/ML community focused on improving user interface design.