Blinkd

Unlock the knowledge in any video, instantly. Blinkd turns lengthy YouTube videos into concise summaries and ready-to-share content. Stop watching, start knowing.

Comment

Inspiration

The idea for Blinkd was born out of a personal frustration. I was watching an insightful video by Alex Hormozi and was captivated by the valuable points he was making. Ten minutes after the video ended, I realized I could barely recall the key takeaways. It struck me that we are surrounded by a wealth of knowledge on platforms like YouTube, but we lack effective tools to capture and retain it. I wanted to build a solution that not only saves time but fundamentally changes how we interact with video content, turning passive viewing into active knowledge acquisition.

What it does

Blinkd is an AI-powered SaaS platform that transforms long-form YouTube videos into structured, actionable summaries. At its core, it solves the problem of information overload.

Here's what it does:

  1. AI-Powered Summarization: Users paste a YouTube URL, and Blinkd's AI (powered by Google Gemini) analyzes the entire video to produce a comprehensive summary, complete with an introduction, titled key insights, conclusions, a closing thought, and relevant tags.
  2. Social Content Generation: For Pro users, Blinkd takes it a step further by using the summary to generate platform-optimized social media content, such as X threads and LinkedIn posts, in seconds.
  3. Knowledge-Sharing Platform: It's more than a tool—it's a community. Users can discover public summaries, follow other creators, and bookmark content. They can also create and share curated lists of summaries on specific topics, making it a collaborative knowledge hub.
  4. Growth & Monetization: Blinkd operates on a freemium model with a built-in referral system. Free users have a generous limit on summaries and bookmarks, while Pro users get unlimited access. Referring friends earns users free credits and Pro users additional free months, creating a viral loop for growth.

How I built it

Blinkd is a full-stack application built with a modern, scalable tech stack, integrating several powerful services to create a cohesive product.

  • Frontend: A hybrid-stack approach using React and Vite. Components are built with JavaScript (JSX) for rapid prototyping, while TypeScript is used for custom hooks and configuration to ensure type safety. Tailwind CSS provides utility-first styling.
  • Backend AI Service: A Python Flask microservice deployed on Google Cloud Run. This serverless architecture was chosen for its scalability, handling the intensive AI processing tasks.
  • AI Engine: Google Gemini 2.5 Pro is the heart of Blinkd. Its powerful multimodal capabilities allow it to process video content directly, going beyond simple transcription to understand context, structure, and key ideas.
  • Database & Authentication: Supabase serves as the backend-as-a-service. It handles the PostgreSQL database, user authentication, and real-time features. I implemented comprehensive Row Level Security (RLS) policies to ensure data privacy and security in this multi-tenant application.
  • Payments: Stripe is integrated for secure and reliable subscription management, handling the logic for the Pro plan.

Challenges I ran into

Building Blinkd was a fantastic learning experience, but it came with its challenges:

  1. Google Cloud Deployment: This was my first time deploying a service to Google Cloud. The biggest hurdle was understanding the entire ecosystem, from containerizing the Python application with Docker to configuring the cloudbuild.yaml for automated builds and deployments on Cloud Run. It was a steep learning curve, but I now have a scalable, serverless backend to show for it.
  2. Complex Authentication State: A persistent bug appeared where the application would get stuck in an infinite loading state if a logged-in user switched browser tabs and then returned. This forced me to do a deep-dive into Supabase's session management. I solved it by implementing robust state tracking in the React AuthContext, handling token refreshes, and using browser visibilitychange and focus event listeners to intelligently re-initialize the user's session, ensuring a seamless and bug-free user experience.

Accomplishments that I am proud of

  • Building a Full-Stack SaaS App: Completing an end-to-end, multi-service application with a viable business model from scratch within the hackathon timeframe.
  • Sophisticated AI Integration: Moving beyond basic API calls to implement a powerful, multimodal AI feature that directly processes video content to provide real, structured value.
  • Solving Complex Technical Problems: Overcoming the challenges with Google Cloud deployment and the nuanced authentication state issues. It proved my ability to learn quickly and engineer robust solutions.
  • Product & Business Vision: Designing not just a tool, but a platform with social features, a freemium model, and a viral referral loop. This shows a holistic understanding of what it takes to build a successful product.

What I learned

This hackathon was an incredible growth experience. I learned:

  • How to deploy, manage, and scale serverless applications on Google Cloud Run.
  • The intricacies of modern web authentication in Single-Page Applications, particularly handling session state and token refreshes.
  • How to leverage the powerful multimodal capabilities of Google Gemini for complex analysis tasks beyond simple text generation.
  • How to structure a secure, multi-tenant database using Supabase and its Row Level Security features.

What's next for Blinkd

Blinkd has a clear roadmap to become a comprehensive content repurposing and discovery suite. Here are the next steps:

  • Enhanced AI Control & Trust: Give users control over the summarization prompts (a premium feature) and integrate AI-powered, real-time fact-checking for YouTube videos to ensure accuracy.
  • Deeper Content Interaction: A "Chat with Summary" feature will allow premium users to ask questions and get answers sourced directly from the video's transcript and metadata.
  • Expanded Content Formats: Allow paid users to create AI-generated audio summaries and short-form videos (for Reels, TikTok, etc.) using AI avatars and custom branding.
  • Creator Economy & Discovery Engine: Transform Blinkd into a discovery platform. We will add analytics to track clicks from summaries to the original YouTube videos. This lays the groundwork for a two-sided marketplace where video creators can reward Blinkd users for driving traffic, and premium users can monetize their high-quality, enhanced summaries.

Built With

Share this project:

Updates