Synapse

Window in the top right corner
Mind Map

Inspiration 🧠💡

We created Synapse out of frustration with the time-consuming process of finding specific information in lengthy videos. As researchers and learners, we've all experienced knowing that critical information exists somewhere in a video lecture or tutorial but being unable to pinpoint its exact location. This problem is especially prevalent in academic settings where valuable insights are buried within hours of footage. We wanted to build a tool that would make video content as easily searchable and accessible as text, saving valuable research time.

What it does 🔍🎯

Synapse is a Chrome extension that uses AI to analyze YouTube video transcripts and connect users directly to the exact moments they're looking for. Users can either search directly through the extension or select text on any webpage and use the context menu to find relevant video content. The extension uses the YouTube API to fetch video transcripts and Gemini's advanced language processing to identify the most relevant segments based on user queries. It then presents these moments with direct timestamp links, allowing users to jump straight to the information they need without wasting time on manual searching.

How we built it 🛠️👨‍💻

🖥️ Built as a Chrome extension using JavaScript, HTML, and CSS
🎬 Leveraged YouTube Transcript API for extracting video content
🤖 Integrated Gemini API for powerful semantic analysis
🖱️ Implemented context menu for right-click searching from any webpage
⚡ Optimized transcript processing for speed and relevance
🎨 Created a clean, intuitive user interface for accessibility

Challenges we ran into ⚠️🚧

We frequently ran into API usage limitations during development, which significantly slowed our progress and forced us to implement creative workarounds. Given the 6-hour time constraint and such a broad topic, we spent a significant portion of our limited time deciding how to approach our app and which specific branch of automation to focus on. Working with unformatted YouTube transcript data lacking proper structure made semantic analysis particularly challenging.

Accomplishments that we're proud of 🏆✨

🔮 Developed semantic search that understands context, not just keywords
🔄 Created seamless Chrome integration that feels natural to use
⏱️ Saved researchers an average of 15 minutes per video in testing
🎯 Built an intuitive interface making AI technology accessible to everyone
🕒 Successfully completed a complex project within a tight 6-hour timeframe

What we learned 📚🧩

The tight 6-hour timeframe taught us the importance of quick decision-making when scoping projects under strict constraints. We gained valuable experience integrating multiple APIs into a cohesive product and developing techniques for working with unstructured data like video transcripts. Perhaps most importantly, we learned to balance technical capability with user experience, ensuring our powerful tool remains accessible despite the complex technology working behind the scenes.