Dub3.ai

Screenshots

Problem Statement

Our main objective is to empower users to effortlessly upload and monetize their voice samples. These samples undergo advanced processing through the Coqui/TTS library by Hugging Face. The magic unfolds within the Cartesi environment, ensuring seamless, efficient, and reliable execution. Once synthesized, the resulting voice is securely uploaded to IPFS (InterPlanetary File System). Simultaneously, an exclusive NFT is minted, representing this synthesized voice as a one-of-a-kind digital asset.

Solution

This project utilizes Cartesi to establish an execution environment for LLM inference. We've chosen to employ coqui/tts (https://huggingface.co/coqui/XTTS-v2) to synthesize input samples uploaded via the dApp.The uploaded sample finds its home on LightHouse Storage within IPFS, following which an NFT is generated containing all necessary metadata.Our NFT contracts are deployed on Polygon Mumbai and Scroll. However, thanks to Chainlink CCIP facilitating cross-chain transactions, the NFT can also be minted on other chains.For development, we employed ScaffoldETH and NextJS for the frontend.