The problem Tapster solves

Tapster addresses the pressing need for high-quality datasets in AI development. Inspired by the story of Mata Shabri, who carefully selected the best fruits for Lord Rama, Tapster aims to create modern-day “Shabris” who generate valuable, classified data for AI models. Users engage with short-form content, like short-videos , Images and games, while selecting between pairs of AI-generated images. Their choices help build high-quality datasets, which AI companies use to improve their models. Built on the Sui blockchain, Tapster ensures data integrity and transparency through attestations, rewarding users for their contributions. By combining entertainment with data generation, Tapster transforms dataset labeling into a fun, gamified experience, addressing the critical challenge of acquiring clean, reliable data.

Challenges we ran into

We faced several challenges during Tapster’s development. Sui’s lack of native support for attestations required us to create a custom SNS attestations contract, which added significant complexity. We also dealt with image persistence issues in Walrus, necessitating extensive debugging. As beginners with the Sui blockchain, writing and deploying smart contracts was a steep learning curve. Additionally, time constraints prevented us from expanding Tapster to handle image-to-text, audio, and video data, though these features are part of our future plans.We faced several challenges during Tapster’s development. Sui’s lack of native support for attestations required us to create a custom SNS attestations contract, which added significant complexity. We also dealt with image persistence issues in Walrus, necessitating extensive debugging. As beginners with the Sui blockchain, writing and deploying smart contracts was a steep learning curve. Additionally, time constraints prevented us from expanding Tapster to handle image-to-text, audio, and video data, though these features are part of our future plans.