Website Visual Optimizer is an AI-powered CLI and library that automates visual audits of web pages. It captures full-page screenshots, applies computer vision to detect inconsistencies in spacing, alignment, and color contrast, then uses a language model to generate detailed design recommendations and CSS diffs. Teams can integrate it into CI/CD pipelines to enforce style guidelines, speed up design reviews, and ensure accessibility compliance, reducing manual effort and improving overall user experience.

Browser Agent integrates OpenAI’s language models with Playwright to perform automated browsing tasks directed by natural language commands. It loads web pages, navigates links, clicks buttons, fills and submits forms, extracts structured data, captures screenshots, and evaluates custom JavaScript. By interpreting GPT output into browser actions, developers can prototype web automation workflows with minimal code. It supports multi-page sessions, cookie and session management, and error handling. Teams can script tasks such as data scraping, end-to-end testing, or dynamic content interaction, all triggered by conversational prompts. Its architecture is modular, exposing hooks for extending capabilities and integrating with downstream processing pipelines.