Noodl.

Why google when you can Noodl.

Comment

Inspiration

It all started with a simple thought. Imagine you’re on a road trip with friends. Everyone’s vibing, and there’s that one person in the group who seems to know something about everything—philosophy, coding, cooking, space-time, you name it. You're impressed, but you also wonder, "How do they learn all this?"

We realized that while the internet gives us access to infinite information, it's often overwhelming. Googling a new topic results in a wall of text and a sea of links. AI chatbots give answers, but they don't provide a structured way to learn. We wanted to fix that. We were inspired to build a tool that makes learning any topic accessible, engaging, and structured, transforming raw curiosity into verifiable knowledge. That's how Noodl was born.

What it does

Noodl is a next-generation, AI-driven educational platform that gamifies your journey through knowledge. At its core, it does three things: Learn, Prove, and Own.

  • Learn: You can type in any topic you're curious about, from "Quantum Physics" to "how to bake a pineapple cake." Our AI, powered by Google's Gemini models, instantly designs a complete learning path for you. It creates an engaging title, a compelling description, and a structured curriculum broken down into bite-sized chapters. Each chapter consists of easy-to-digest slides and interactive quizzes to reinforce your understanding. For those moments of spontaneous curiosity, there's even an "I'm Feeling Lucky" button that suggests novel topics to explore.

  • Prove: As you complete each learning path, your achievement isn't just a checkmark in an app. You are awarded a unique NFT certificate, minted directly to your Web3 wallet. This certificate features a unique, AI-generated pixel art emblem symbolic of the topic you mastered. It serves as permanent, verifiable proof of your achievement that you truly own.

  • Own: Every learning path you create is registered on the Ethereum (Sepolia Testnet) blockchain with a unique content hash, providing an immutable, tamper-proof record of the curriculum. Your NFT certificates are ERC721 tokens that belong to you, which you can view in your Noodl profile or in any NFT-compatible wallet like MetaMask.

To better understand in depth, see how Parth has explained in his YouTube video: https://youtu.be/u6OgVPZA8H4

How we built it

Noodl is a full-stack application built with a modern, robust architecture.

  • Frontend (Mobile App): We used Flutter to build a cross-platform mobile app with an engaging, gamified user interface. State management is handled by the Provider package, and we ensured a seamless user experience by persisting wallet sessions using SharedPreferences. For secure and seamless Web3 login, we integrated WalletConnect v2 and the MetaMask SDK.

  • Backend (API & Services): The server is a Python application built on the Flask web framework. It communicates with Google's Gemini API for all AI-driven tasks: dynamic course creation, semantic search embeddings, and generating the unique pixel art for NFT certificates.

  • Infrastructure & Web3: Our entire database, including user data, learning paths, and progress logs, is powered by Supabase, where we also utilize its vector (pgvector) extension for our state-of-the-art hybrid search. The backend interacts with the Ethereum blockchain using Web3.py. All NFT assets (images and metadata) are stored decentrally on IPFS via Pinata. The on-chain logic is handled by Solidity smart contracts for both the Learning Path Registry and the ERC721 NFT Certificates.

How about you try our app by following this link which leads to the GitHub Release section of our .apk file: https://github.com/AyushDhimann/Noodl/releases/tag/stable

Challenges we ran into

Building a full-stack, AI- and Web3-integrated application came with its share of challenges.

  1. Reliability of AI and Blockchain Operations: AI generation and blockchain transactions can be long-running and occasionally fail. To handle this, we architected the backend to run these processes in asynchronous background threads. This provides an immediate response to the user, who can then poll for status updates. We also implemented robust retry mechanisms for API calls and designed our smart contracts to be idempotent, making the system resilient against desynchronization between our database and the blockchain.
  2. Platform-Specific Web3 Integration: We encountered a significant hurdle with MetaMask's WebView connection being blocked on Android 15 and newer. We solved this by implementing a manual wallet address login as an alternative, ensuring users on all platforms have a smooth onboarding experience.
  3. Content Structuring: Getting an AI to produce not just text, but a well-structured, multi-level course with interleaved slides and quizzes, required sophisticated prompt engineering and a multi-step generation process. We designed the system to be atomic, so if any step fails, it gracefully rolls back to prevent partially-created, unusable content in the database.

Accomplishments that we're proud of

We are incredibly proud of creating a fully-functional, end-to-end platform that seamlessly merges AI, mobile development, and Web3 technology. The AI-powered course generation works dynamically for any topic, and the on-chain credentialing provides real, tangible ownership to the user.

We're particularly proud of the user experience. From the smooth, gamified learning interface and the "I'm feeling lucky" discovery feature to the AI-generated audio for slides, we focused on making learning fun and accessible. The fact that every course's curriculum and every user's certificate is verifiably on-chain is a huge accomplishment that realizes our vision of decentralized, provable education.

What we learned

This hackathon was a massive learning experience. We dove deep into the Google Gemini API, learning to harness its different models for text, vision, and embeddings. On the Web3 side, we gained hands-on experience with Solidity, Web3.py, IPFS, and the complexities of WalletConnect v2 integration in a Flutter app. Architecting the asynchronous, resilient backend taught us invaluable lessons about building robust, real-world systems. Above all, we learned how to bring together these disparate, cutting-edge technologies into a single, cohesive, and user-friendly product.

What's next for Noodl

This is just the beginning for Noodl. We envision a future where it becomes a thriving, decentralized ecosystem for learning. Our roadmap includes:

  • Enhanced AI and Personalization:

    • Score Analysis & Adaptive Learning: The app will track user performance in quizzes and automatically suggest revisiting specific topics or concepts where they performed poorly, creating a truly personalized learning loop.
    • Voice Assistant: We plan to integrate a full voice assistant, allowing users to navigate the app, generate Noodls, and listen to lessons entirely hands-free.

  • Improved Discovery and Accessibility:

    • Advanced Search & Filtering: Users will be able to search for Noodls and filter them by category, difficulty level, or estimated completion time.
    • Light Mode: To complement our current dark mode, we will introduce a light mode for improved user comfort and accessibility.

  • Deeper Gamification and Social Interaction:

    • Leaderboards and Badges: We will introduce leaderboards to foster friendly competition and a wider array of badges to celebrate learning milestones.
    • Community Q&A: A dedicated section for each Noodl where users can ask questions, share insights, and help one another learn.

Built With

  • AI: Google Gemini API (text-embedding-004, gemini-pro, gemini-pro-vision)
  • Frontend: Flutter, Provider, WalletConnect v2 (web3modal_flutter), MetaMask SDK
  • Backend: Python, Flask, Web3.py
  • Database: Supabase, PostgreSQL, pgvector
  • Blockchain: Solidity, Ethereum (Sepolia Testnet)
  • Decentralized Storage: IPFS, Pinata
  • Demo UI: Gradio

Built With

Share this project:

Updates