NodeSurfer

• 

  planning document

Inspiration

We were inspired by how the human brain naturally connects ideas. It does so not in rigid outlines, but in flexible webs of meaning, all connected to one another. Traditional knowledge-intake tools, like Anki (flashcarding) and Quizlet (Multiple Choice Quizzes), fail to reflect this. We want to create a smarter, more intuitive way for students and learners to visualize, explore, and understand complex topics—just like how the mind actually does. The future of learning lies in user interactive apps that can adapt well to a given user, not in the one-to-one flashcarding of the past.

What it does

NodeSurfer lets users upload notes (PDF, DOCX, TXT, Markdown, etc.) and automatically extracts key concepts to generate a dynamic knowledge graph. This graph shows how ideas connect, forming a visual map of the topic. Users can then chat with an AI assistant that’s aware of their nodes and the concept relationships, helping them explore, clarify, and understand anything in their material.

How we built it

We used a combination of: text extraction libraries to handle different file formats (PDF, DOCX, TXT, etc.) and convert them to Markdown, OpenAI tools for summarizing and extracting key concepts into nodes, CedarOS for the chatbot and front-end, and the Mastra framework to tie different frameworks together. We tied it all together in a responsive web interface that features a fully implemented backend in Python and PostgreSQL.

Challenges we ran into

• Combining the front end, backend, and the APIs.
• Deploying the website, as the free GoDaddy website was not quite available.
• Making the visuals on the knowledge graph easy to access and easy to use. ## Accomplishments that we're proud of
• Successfully built a working pipeline from raw files to an interactive concept map!
• Implemented a chatbot that can reference specific nodes and explain things in context to the user.
• Support for users built-in, with nodes being able to be saved on each workspace file. ## What we learned
• Backend and Frontend construction, as well as how to assemble a full stack of APIs, frameworks, and languages to make everything come together in one web app.
• Good UX is critical when dealing with complex and interactive data visualization. ## What's next for NodeSurfer
• Add support for handwritten notes and lecture audio transcription.
• Improve personalization so the graph and chatbot evolves with each user’s learning journey.
• Enable collaboration, allowing users to share and build on each other's concept maps.
• Integrate spaced repetition and quiz generation based on the knowledge graph, to better help them lock in on concepts.
• Launch a mobile version for on-the-go learning and note interaction!

Built With

• cedaros
• css3
• fastapi
• godaddy
• html5
• javascript
• mastra
• next.js
• openai
• postgresql
• pydantic
• python
• react
• sql