📚You're viewing a previous hackathon event (Nov 2025)
NOVEMBER 8-9, 2025

HackASU

24 hours. Unlimited creativity. Build with Claude AI at ASU's most innovative hackathon.

Memorial Union, Pima (230)

11 AM Start • Nov 8-9

HackASU 2025 Winners

Congratulations to all the incredible teams who built amazing projects!

🏆 1st Place

GamEd.AI

GamEd.AI

"Play it. Learn it. Master it."

GamEd.AI converts any text — a paragraph, a concept, or a question — into an interactive mini-game that teaches through play.

It builds visual stories, puzzles, and reasoning challenges that guide students to understand concepts, not just recall them. Each click, drag, or decision becomes a signal of mastery — making it both a learning platform and an assessment engine.

Key Features:

  • Text-to-game engine using NLP to extract key learning concepts
  • Reasoning graph that converts ideas into interactive story nodes and puzzles
  • One-click "Generate Game" interface from raw text
  • Auto-reasoning path tracker that visualizes student logic
  • AI-proof tasks requiring reasoning, not memorization

Technologies used: Claude API, FastAPI, React, Canvas/WebGL, OpenAI, LangChain, PostgreSQL, Redis, TypeScript

TEAM

Mayank VyasYash Sanjiv ShahShiven AgarwalPriyanuj BordoloiAshish Raj Shekhar
🥈 2nd Place

Let's Lock In

Let's Lock In

"An ADHD-smart study platform that makes focus feel like a game, not a fight."

Let's Lock In syncs with Canvas LMS to transform overwhelming coursework into manageable, motivating tasks for students with ADHD.

Around 15–17% of college students report ADHD symptoms. This platform was built for how ADHD brains actually work — turning structure, focus, and organization into something engaging, rewarding, and supportive.

Key Features:

  • Fetches assignments, due dates, and effort estimates from Canvas
  • Gamified streaks, badges, and progress bars to build momentum
  • Pomodoro-style work chunking for focused sessions
  • Optional attention checks via face/pose estimation (MediaPipe)
  • In-app notes, summaries, and study cheatsheets

Technologies used: React.js, Node.js, Express, PostgreSQL, OAuth2, MediaPipe

TEAM

Himanshu Chowdary Kollaanushka TiwariChaitanya Nookalavishnu uppalapati
🥉 3rd Place

Impulsa

Impulsa

"Making the journey of studying and working abroad easier."

Impulsa is an AI-powered platform that helps students and professionals adapt to life abroad, focusing on the emotional and social sides of culture shock.

The app combines interactive learning sections (inspired by Duolingo's level-based approach) with AI chat powered by Claude and GPT. Users can practice cultural scenarios, ask questions, and progress through levels designed to improve real-world understanding.

Key Features:

  • Level-based cultural learning modules
  • Claude-powered AI chat for cultural Q&A
  • Demo forum for community insight sharing
  • Polished, globally-accessible UI

Technologies used: React, TypeScript, Next.js, Tailwind CSS, Clerk, Claude API

TEAM

Matthew CruzLakshya DhingraJesus Aguilar
🏆 1st Place

VisionForge

VisionForge

"Instantly design powerful neural networks with zero code."

VisionForge is a no-code development environment that lets users visually design, replicate, and experiment with deep learning architectures — then export clean, production-ready PyTorch or TensorFlow code in minutes.

While many no-code tools exist for agents, VisionForge fills a clear gap: designing the core deep learning architectures that power those agents.

Key Features:

  • Drag-and-drop canvas for building full ML pipelines
  • HuggingFace model import for hybrid/multimodal experiments
  • Claude AI assistant for natural language architecture modification
  • One-click PyTorch / TensorFlow code export
  • Real-time tensor shape inference across branching and skip connections

Technologies used: React 19, TypeScript, Vite, Tailwind CSS, ReactFlow, Django, Claude AI, Zustand, Framer Motion

TEAM

Aaditya JindalBhavya ShahAakash KheparGunbir Singh
🥈 2nd Place

Crodex

Crodex

"Switch stacks without rewriting — AI-powered framework conversion."

CrodeX converts code between frameworks using AI, preserving logic and functionality so development teams can switch stacks without losing the value in existing code.

It reduces migration effort from weeks to hours, demonstrating how AI can automate technical tasks while preserving business logic.

Key Features:

  • Detects frameworks from file patterns and dependencies
  • AI-powered conversion using Anthropic and Gemini APIs
  • Supports Flask, Django, Spring Boot, and more
  • Real-time progress tracking via Server-Sent Events
  • Session-based project state management

Technologies used: Flask, Python, Anthropic API, Google Gemini API, Redis, HTML5, CSS3, JavaScript ES6+

TEAM

Abhinav RejaKeyur Ashok MadaneSumedh GajbhiyeUkani Vedant
🥉 3rd Place

unbored

unbored

"Take the boredom out of onboarding."

unbored turns any GitHub repository into a polished onboarding site in minutes — giving new contributors a "first hour" experience without weeks of reading scattered READMEs and closed PRs.

New contributors routinely spend 2–4 weeks piecing together tribal knowledge. We think there's a better way.

Key Features:

  • AI-generated architecture overview summarizing services, modules, and data flow
  • Setup guide with pitfalls mined from historical issues and PRs
  • Interactive file dependency graph (Tree-sitter import parsing)
  • FAQ auto-generated from recurring questions in repo history
  • Exports a static Docusaurus site ready for GH Pages / Netlify / Vercel

Technologies used: Python, Typer CLI, Claude API, AWS Bedrock, Lambda, Docusaurus, Node.js, TypeScript, React

TEAM

Pranjal PadakannayaAkash BagchiAkshaya NadathurSachin Shivanand Shankarikoppa
📊 1st Place — Polymarket x Claude

Kairos

Kairos

"A personalized one-stop agentic workflow for prediction markets."

Kairos is a "markets mission control" that syncs Polymarket, Nevua, Adjacent, and Claude to find mispricing, manage alerts, and prove outcomes — all in one coherent workspace.

A Claude-powered copilot with persistent memory stores your trades and rules, then anticipates your next move.

Key Features:

  • Claude-powered agent with persistent trade memory (Postgres JSONB)
  • Coherence module: streams live Polymarket prices, encodes overlapping markets as a boolean matrix, and runs feasibility checks
  • Nevua integration for server-side watchlists and live alerts
  • Adjacent API "Newsroom Mode" pairing market movers with journalism
  • Factor graph + message passing for interpretable probability beliefs
  • Arena module that red-teams resolution criteria for exploits

Technologies used: Next.js 15, React 19, NextAuth 5, Prisma, Neon/Postgres, Tailwind v4, Radix UI, Recharts, TypeScript

TEAM

Aashir Javed
🥈 2nd Place — Polymarket x Claude

gibbsAlpha

gibbsAlpha

"Market data turned into real insight via a multi-agent reasoning pipeline."

gibbsAlpha converts raw market data into structured, reliable insight using an EMNLP-grade multi-agent reasoning pipeline directly inspired by recent research.

Instead of a black-box chatbot, it behaves like a coordinated research team.

Agent Pipeline:

  • Analyzer Agent — statistical checks, conditional probabilities, and expected value reasoning
  • Context Agent — narrative structure from market history and cross-asset signals
  • Risk Agent — uncertainty, volatility, and fragility evaluation
  • Synthesizer Agent — merges partial analyses into one clear, human-readable explanation

Technologies used: Claude 3.5 Sonnet, n8n orchestration, Next.js 14, Node.js/Express, Supabase, Polymarket API, WebSockets, TypeScript, Tailwind

TEAM

Utkarsh ByahutMuhammed Topiwala
⛓️ 1st Place — EtherFi x Claude

GitAccountable

GitAccountable

"Stake crypto, commit code daily, or lose it all."

GitAccountable is a crypto-accountability app where you put real skin in the game: stake 0.01 ETH, commit to 7 days of GitHub activity, and either claim your stake plus staking rewards — or forfeit everything if you miss a day.

A Chainlink oracle verifies your commits daily. No cheating.

Key Features:

  • Solidity smart contract (Foundry) integrating mock Ether.Fi liquid staking
  • Chainlink Functions oracle querying the GitHub API for daily commit verification
  • React frontend with Wagmi/RainbowKit for Web3 interactions
  • Mock liquid staking rewards working alongside smart contract accountability logic

Technologies used: Solidity, Foundry, Chainlink, Ether.Fi, React, Wagmi, RainbowKit, Node.js, GitHub API

TEAM

John GJacob Blemaster
🥈 2nd Place — EtherFi x Claude

Yield Quest

Yield Quest

"A no-loss prediction game powered by EtherFi weETH with Claude AI as your Game Master."

Yield Quest transforms DeFi engagement by letting you play prediction markets using only the yield generated from your staked weETH — your principal stays 100% safe and withdrawable at any time.

Think of it as a gamified yield layer on top of EtherFi's liquid staking.

Key Features:

  • Deposit weETH → earn Yield Credits (YC) → bet only YC on binary YES/NO markets
  • Claude AI (Haiku 4.5) as Game Master: personalized quests, probability hints, and automated trading mode
  • 7 achievements with rarity tiers (Common → Legendary)
  • Global leaderboards: Accuracy, Wisdom Index, and Quest Masters
  • Wisdom Index = composite score (accuracy 50% + yield efficiency 30% + streak bonus 20%)

Technologies used: Next.js, Solidity, React, Prisma, Anthropic SDK, RainbowKit, Wagmi, Tailwind CSS, TypeScript

TEAM

Ashwanth Balakrishnan
🥉 3rd Place — EtherFi x Claude

E.V.O.

E.V.O.

"Smart Options. Instant Premiums."

E.V.O. (Ether.fi Vault Options) is a full-stack decentralized application allowing eETH holders to earn additional premium income on their assets through two core collateralized strategies.

Strategies:

  • Covered Calls: Lock eETH as collateral, sell call options, earn instant USDT premium
  • Cash-Secured Puts: Lock USDT, sell put options, earn instant USDT premium

Key Features:

  • Instant premium payment at collateral lock time (not at expiry)
  • Claude API integration for plain-English trade risk/reward analysis
  • Real-time APR calculation engine via Node.js backend
  • Live market data via Tatum.io integration
  • One-command local deploy: contracts, test wallets, and servers

Technologies used: Solidity, Hardhat, Next.js 14, React, Wagmi, Viem, RainbowKit, Node.js, TypeScript, Claude API

TEAM

Aditya KumarDhanush KoyiDhyan Sanjaykumar Patel
📍

Check-In & Room Information

Follow these steps when you arrive

1

Check-In at Turquoise room in MU (Room #220)

Start by checking in at the Turquoise desk to get verified

2

Main Room: MU Pima (230)

After check-in, proceed to Memorial Union Pima room for the main event

3

Overflow: LSE 104

If MU Pima is at capacity, overflow participants will be directed to LSE 104

⚠️ Important: You must check in at Turquoise before entering any hackathon rooms. Verification is required for entry.

Hackathon Tracks

Choose your challenge and compete for track-specific prizes

📚

Education & Social Good

Create tools that make learning more engaging, accessible, or personalized. Or create solutions that address social challenges.

🛠️

Developer Tools

Create tools that enhance developer productivity, collaboration, or code quality using Claude API.

💰 Sponsored
⛓️

EtherFi x Claude

Build innovative solutions combining EtherFi's liquid staking protocol with Claude's AI capabilities. Create tools for DeFi analytics, portfolio management, staking optimization, or educational resources about liquid staking.

💰 Sponsored
📊

Polymarket x Claude

Leverage Polymarket's prediction markets with Claude AI to build market analysis tools, trading assistants, event outcome analyzers, or educational platforms that make prediction markets more accessible.

24-Hour Schedule

Day 1 - November 8

11:00 AM - 11:30 AM

Registration & Welcome

Check-in, networking, pictures, and a speech by Ben Zhou (NLP Professor at ASU)

11:30 AM - 12:00 PM

Opening Ceremony

Welcome from organizers and Anthropic sponsors. Hackathon rules, prizes, track announcements, and Claude API demo

12:00 PM - 1:00 PM

Hacking Begins! Team Formation & Ideation

Teams finalize and register their projects. Access to mentors for initial concept validation. Pizza available from Hungry Howie's

1:00 PM - 8:00 PM

Hacking Sprint

Mentors circulating, help desk for API issues, track-specific mentor hours

8:00 PM - 9:00 PM

Minecraft Tournament Break! 🎮

Optional team-based Minecraft mini-games tournament. Prizes for winners. Relaxation area for those who prefer to rest

9:00 PM

Go Home and Rest!

Remember to get a good night's sleep, or take advantage of it to hack overnight!

Day 2 - November 9

9:00 AM - 11:00 AM

Final Sprint

Last push for features and polish. Code freeze reminder at 9:30 AM

11:00 AM - 11:30 AM

Submission Deadline & Setup

Hard deadline for submissions. Teams prepare demos

12:30 PM - 1:30 PM

Lunch Break

Costco Pizza lunch break

3:00 PM - 4:00 PM

Closing Ceremony

Winner announcements for each track + grand prize. Sponsor remarks. Distribution of prizes. Closing remarks

4:00 PM

Event Ends

See you next year!

Our Sponsors

Thank you to our amazing sponsors who make this event possible

Platinum Sponsors

Partners

Hackathon Rules

Keep it fair, keep it fun, keep it innovative

1
👥

Team Requirements

Teams can have 1-5 members. Solo hackers are welcome!

2
🎓

ASU Student Eligibility

ASU students are eligible for all prizes (API credits, cash, prizes, and merch). Must complete registration form in-person at the hackathon on Saturday or Sunday.

3
🌐

Non-ASU Student Eligibility

Non-ASU students are only eligible for Anthropic merch prizes. If your team includes one non-ASU student, they will not be eligible for other prizes.

4
💻

ASU Online Students

Must provide verifiable proof of ASU Online enrollment. With verification, eligible for all prize categories without in-person form completion.

5

Work During Event Hours

All projects must be created during official Hackathon hours. Exceptions: publicly available open-source code and standard libraries (must comply with licenses).

6
📝

Academic Integrity

No plagiarism. You must disclose and properly attribute all APIs, libraries, third-party resources, or pre-written code used in your project.

7
🚫

Double Submission Prohibited

No submitting the same project to multiple tracks or hackathons. Each project must be unique to this event.

8

Demos should be Functional

Any project or application submitted must be able to be verified by judges. Non-functional demos will be sent to further review.

9
📹

Demo Video Required

Submit a 3-5 minute demo video showing your project in action. Judging will be based on video submissions.

10
🤝

Respect the Venue, Organizers, and Participants

Treat the hackathon space with respect. Follow instructions from organizers and staff. Maintain a positive and collaborative environment for all participants.

11
📜

Code of Conduct

Harassment of any kind (based on gender, race, age, sexual orientation, disability, nationality, or religion) will not be tolerated and will result in removal from the event.

12
📜

Compliance with Laws

All participants must comply with local, state, and federal laws while participating in the hackathon.

Meet Our Judges

Industry experts who will evaluate your projects

Sean Yen photo

Sean Yen

Judge

Software Engineer @ Tesla

Dylan Lu photo

Dylan Lu

Judge

First Place Winner @ CalHacks

Ben Zhou photo

Ben Zhou

Judge

ASU Assistant Professor @ SCAI

Ready to Build?

Join us for 24 hours of innovation, collaboration, and creativity. All skill levels welcome!

ASU Claude Builder Club