ml5.js is an open-source JavaScript library that brings machine learning into the browser, explicitly aimed at being “approachable for a broad audience of artists, creative coders, and students” rather than only ML engineers. It’s built on TensorFlow.js and ships pre-built models for pose estimation (body, hand, and facial landmarks), image and sound classification, and custom neural network training.
Making browser-based ML approachable for creative coding is the real contribution – most ML tooling assumes a Python data-science background, while ml5.js targets people who already think in JavaScript and want to add gesture control, sound-reactive visuals, or classification to a creative-coding project without a separate ML pipeline. Development is led by NYU’s ITP/IMA and NYU Shanghai IMA programs, with active community contributions.
This is a code library, not a hosted product – no signup or dashboard, just an import into your own JavaScript project. It’s free and open-source, with comprehensive documentation and learning resources for people new to both ML and creative coding.
