Inspiration We live in an era of information overload where we consume more than we can process. I realized that our brains are great for having ideas, but not for storing them. I wanted to build "Cortex" to act as a digital neocortex—a seamless layer of intelligence that captures, organizes, and retrieves information exactly when you need it, effectively ending the "tab-hoarding" and "forgotten bookmarks" era.
What it does Cortex is an AI-powered personal knowledge companion that functions as a "Second Brain." It doesn't just store links or notes; it understands them. Using advanced RAG (Retrieval-Augmented Generation), Cortex allows users to chat with their saved content, discover hidden connections between different topics, and transform scattered bookmarks into a structured, searchable knowledge base. It’s built to reduce cognitive load and maximize creative output.
How we built it The core of Cortex is built using Python and FastAPI for a high-performance backend. We utilized Gemini models and vector databases to enable semantic search and contextual understanding. For the interface, we focused on a minimalist and modern design (Kotlin) to ensure the user stays in "flow state." The integration layer leverages Azure for scalable AI processing.
Challenges we ran into One of the biggest hurdles was managing "context window" limitations while ensuring the AI had access to the user's entire knowledge base. We solved this by implementing an efficient embedding and chunking strategy. Additionally, designing a cross-platform experience that feels like a natural extension of the user's workflow—rather than just another app to manage—required significant UI/UX iteration.
Accomplishments that we're proud of We are incredibly proud of our semantic retrieval engine, which can find relevant information even if the user doesn't remember the exact keywords. Successfully launching the "Cortex: AI Bookmarker" on the Google Play Store and seeing it handle complex queries across hundreds of saved documents was a major milestone for the project.
What we learned Building Cortex taught us that AI is not just about the model, but about the data architecture behind it. We learned deep insights into Vector Embeddings and how to fine-tune the balance between AI autonomy and user control. On the product side, we realized that the best productivity tools are those that disappear into the background and work silently.
What's next for Cortex: AI Second Brain The vision for Cortex is to become platform-agnostic. Currently we are developing a browser extension and a desktop app to provide a truly ubiquitous experience. Future updates will include "Autonomous Insights," where Cortex proactively suggests connections between your notes, and integration with local LLMs to ensure maximum privacy for personal data.
Log in or sign up for Devpost to join the conversation.