TLDW (Too Long; Didn't Watch) is a powerful tool designed to transform lengthy YouTube videos into an efficient and structured learning experience. This application allows users to effortlessly convert any video URL into digestible highlight reels and timestamped AI-generated answers, facilitating the absorption of content that would otherwise take an hour to view in just a few minutes.

Project Features

• AI Highlight Reels: Featuring two generation modes—Smart for quality and Fast for speed—users can create highlight reels that play all segments or regenerate content based on themes.
• Gemini-Powered Insights: The application provides quick previews, structured summaries, suggested questions, and memorable quotes displayed side by side for easy reference.
• Interactive AI Chat: Engage in a conversational chat based on the video transcript, receiving structured JSON responses with timestamp citations, enhanced even further by fallback options when rate limits are reached.
• Transcript Synchronization: A specialized transcript viewer remains in sync with the YouTube player, allowing users to click any sentence to either navigate or capture quotes effortlessly.
• Personalized Note Workspace: Users have access to a dedicated workspace for taking notes alongside the transcript, chat, and source material. An integrated /all-notes dashboard enables review across different videos.
• Authenticated Library: Users can save their analyses, bookmark favorites, and manage preferences through authenticated library pages, which are powered by Supabase.
• Rate Limiting and Caching: Efficient handling of requests is enabled through aggressive caching of previous analyses and differentiated rate limits for signed-in and anonymous users.
• Robust Security Measures: The application implements strong security protocols, including Content Security Policy (CSP) headers, CSRF protection, and data-rate limiting based on user state.

Underlying Architecture

• Frontend Stack: Built using Next.js 15, React 19, TypeScript, and Tailwind CSS v4, the frontend is designed for a smooth user experience with shadcn/ui components.
• Backend Structure: Utilizing Next.js serverless route handlers augmented by security middleware for input validation and rate limiting, the backend processes AI tasks and handles user interactions.
• AI Processing: The AI features are orchestrated through specialized libraries that handle Gemini 2.5 models with structured output and efficient transcript processing.
• Data Management: Supabase is employed for persistent storage including video analyses, user history, and profile data, ensuring efficient management of content.
• API Surface: The application encompasses a range of API endpoints for video information ingestion, AI generation, user data management, conversational tools, and security utilities.

By streamlining the way users engage with educational videos, TLDW fosters a more productive and enjoyable learning environment, providing invaluable insights at a fraction of the usual time investment.