Do[NFT] is a decentralized protocol offering AI-powered NFT modification. Using the protocol, NFT holders can mint brand new NFTs using their existing NFT assets. The new NFT is inextricably linked with an initial one. The new NFT asset is modified initial NFT created by the application of AI-generated effects such as gif animation, filters, artistic noise, etc. The modification is done by neural network algorithms and the result cannot be exactly predicted. It leads to an opportunity to mint different NFTs, including rare ones that nobody could even expect. Such an approach allows offering a brand new experience and a lot of fun/opportunities to grow their collection to NFT owners. It is a permissionless dApp where anybody can offer a new AI-based filter, animation technique, or face aging algorithm for community usage (for free or charging certain service fees). Our goal is to build a dApp offering to its users a broad variety of NFT filters/modifiers (provided mainly by community contributors) and: 1. Get integrated into leading NFT marketplaces allowing to modify any existing NFT there on-the-fly 2. Become “masks and filters of Instagram” for the NFT world. As nobody could accurately predict the results of modification (it leads to an endless variety of different outcomes) such a service provides a lot of fun based on the Skinner box effect. It means that the NFT community will be engaged in minting, trading, discussing the results, or even creating a whole generative art gallery from a single Bored Ape NFT.

DoNFT showcase

How it's made

We use rarible protocol and opensea api for aggregate user's NFTs. Our own smart contract for bundling d few NFTs to one. We trained a style transfer and segmentation models on Bored Apes NFTs dataset and also we finetuned a image animation model. We deployed our contract on rinkeby Testnet, Harmony testnet and Mumbai Polygon testnet.