Fragment started as an idea in 2017 when I did a VJ performance based on three.js visuals with a friend. We had a couple of different scenes that could react to audio beats and MIDI inputs. All scenes were made at different times, with different stacks and MIDI inputs had to be programmed and reminded beforehand.

In 2018, we did a projection mapping for MAPP Montreal, also made with three.js. Visuals were still reacting to the soundtrack but it was not live, we needed to export a video of the whole thing. As we were lacking time to find better solutions, we ended up using a computer capable of capturing the screen while still rendering at 60fps. Not ideal.

In 2019, I became a freelance creative developer and extended the variety of my practices: web development, real-time graphics, and generative art. I felt the need of a sketching environment, who could support the range of needs of my previous and future projects:

• programmable interface
• reactive to different kinds of inputs (GUIS, keyboard, MIDI, mouse, sound...) from the code and the interface
• fast as possible
• images, videos and live exports

So I spent the last years building my own tool.