DogeRace

technical architecture
figma low fidelity prototype

Credentials

username1 password1

Inspiration

We were inspired by the rapid rise of AI models—and the real challenge users face choosing the right one for their specific needs. Reviews often say “they’re all good,” but don’t help with targeted use cases like stable art generation or accurate stock predictions. We wanted to make comparing AI models transparent, fun, and engaging.

What it does

DogeRace is a Bolt-based game where users bet on which AI model will perform best in real-world prediction challenges. Each race is a head-to-head AI comparison, with results stored immutably on the Algorand blockchain for verifiable, trusted proof. Users earn crypto payouts for backing the winning AI.

How we built it

We used Bolt to design a playful, intuitive betting UI. The backend is built with AWS serverless services—API Gateway, Lambda, and DynamoDB—to handle AI predictions, scoring via Assessor AI, and wallet management. We integrated Algorand for on-chain storage of race results and payouts, ensuring full transparency and trust.

Challenges we ran into

We faced challenges orchestrating multiple AI model calls within Lambda’s time limits, integrating testnet crypto wallet flows, and ensuring seamless communication between the Bolt front-end and our AWS serverless APIs.

Accomplishments that we're proud of

We’re proud of building a fully serverless, blockchain-integrated app that turns AI benchmarking into a fun game. Our design makes blockchain invisible to users while delivering true transparency.

What we learned

We learned the power of combining serverless architecture, blockchain technology, and low-code UI platforms like Bolt to deliver real-world, scalable solutions quickly.

What's next for DogeRace

We plan to expand DogeRace with more AI models, advanced betting markets, richer user profiles, and NFT-based achievements. We also aim to deploy on Algorand mainnet to enable real payouts and grow a community of users crowdsourcing trusted AI comparisons.