Generative art is a distinct form of art that often uses autonomous systems or algorithms to randomly generate content.

According to Tate Modern, the practice has roots in the Dada art movement and was largely pioneered by abstract painter Harold Cohen, who was among the first to use computer-controlled robots to generate paintings in the late 1960s. More recently, it has gained popularity among collectors and creators of non-fungible tokens (NFT).

While generative art is an umbrella term that encompasses many types of styles and mediums, there are common characteristics that have come to define the genre, including the repetition of patterns, shapes, colors and motifs; the randomness of composition; the use of an algorithm to generate images; and the use common use of geometric patterns and shapes.

Put simply, an artist will input a set of rules (for example, a range of colors and patterns) against a number of iterations and randomness. The artist then lets the computer generate the artwork within this framework.

While the art style has been around for several decades, it has recently gained popularity as a tool for creating non-fungible token artwork. In 2014, digital artist Kevin McCoy minted what is considered to be the first known NFT, titled “Quantum,” which used code to create an abstract image that was registered on an early blockchain network.

A number of similar experiments soon followed, including Autoglyphs by Larva Labs, which has been hailed as the first on-chain generative art project built on Ethereum. Today, some generative art NFT projects make use of a blockchain’s smart contract to execute randomized, computer-driven code when certain conditions are met, generating a unique piece of art for the NFT holder once minted.

As the movement has grown over several decades, so has the list of tutorials, tools and resources that are used by new creators and collectors in the NFT space.

Although the overall process depends on the complexity and functionality of the concept, here are several popular tools used by generative art creators:

While most programming languages can be used to create generative art, many artists prefer to use JavaScript and its p5.js creative coding library. Online communities have also sprung up to discuss additional tools that can be added to any generative art toolbox.

Beyond coding, there are also popular resources that use artificial intelligence to help generate art, including:

Additionally, Async Art has become a popular tool for artists or musicians looking to create generative art on the Ethereum blockchain without the use of code. The tool allows NFT creators to upload their assets, set rarity percentages and create an entire generative NFT collection, all in one place.

Released in 2019 by CryptoPunks creator Larva Labs, Autoglyphs are “an experiment in generative art,” each created by code running on the Ethereum blockchain. Anyone could create a glyph by donating a “creation fee” of 2 ETH (about $35 at the time) to Larva Labs’ chosen charity 350.org. The project capped the number of NFTs at 512, shutting down the generator forever. Autoglyphs have also been made into generative music.

Tyler Hobbs is a visual artist from Austin, Texas, who develops and programs custom algorithms that are used to generate visual imagery. He is known for his algorithmically generated Fidenza collection, which produced “unpredictable, organic curves,” with rarity determined by the mix of different-sized elements. He recently released QQL, a collaborative project with Dandelion Wist, the co-founder of generative art platform Archipelago. While anyone can play around with the QQL algorithm, only holders of the mint pass, which sold out on Sept. 28, can turn the artwork into an official NFT in the QQL collection.

Created by the enigmatic persona behind Sotheby’s “The Fungible” collection, Lost Poets is both an NFT collectible and strategy game by digital artist Pak. The collection includes 65,536 obtainable NFTs and 1,024 Origin NFTs. Further details can be found in the project’s cryptic roadmap. According to the project's website, each NFT is not made up of modular pieces but was created by an AI specifically designed for the project.

Created by Erick Calderon, aka Snowfro, as the genesis project of the on-chain Art Blocks platform, Chromie Squiggles are unique, randomly generated squiggles of colors in nine different style schemes.

Created by artist Dmitri Cherniak and generated from a variety of combinations using Javascript, Ringers are a set of 1,000 unique rings based on different combinations of “strings and pegs.”

Robbie Barrat is an artist and graphic designer who works with artificial intelligence. He is known for his use of generative adversarial network (GAN) machine learning models.

As a self-taught programmer, Matt Kane gained recognition in September 2020 with his Async Art piece “Right Place, Right Time” that sold for 262 ETH (approx. $100,800) in September 2020. His most recent collection, Gazers, on ArtBlocks is a series of 1,000 NFTs based on the cycles of the moon, which evolves over time.

Matt DesLauriers is a Canadian artist and coder. His long-form generative art project titled Meridian contains an algorithm that creates stratified landforms using “hundreds of thousands of small strokes of color.”

Most generative art projects can be bought and sold on popular NFT platforms including OpenSea and Rarible, although specialized generative art platforms have recently sprung up, including:

A number of generative art NFT projects have gained widespread recognition, especially those tied to established artists. Specialized platform Art Blocks has generated over $1.3 billion in volume on primary and secondary markets to date, according to DappRadar, highlighting the success of generative art projects.

Notable sales over the past several years include: