---
post_type: "project"
title: "HIRO"
blurb: "A marketplace for generative and interactive artworks — pieces you run and reshape rather than just look at."
tags:
    - SAAS
    - Online platform
    - Creative coding
    - Generative art
    - Interaction design
year:
    - 2021
    - 2024
current: false
link: "https://hir.ooo"
stack:
    - p5.js / creative coding
    - JavaScript SDK
    - Node CLI
    - Real-time editor
    - Payments
---
HIRO is a marketplace for generative and interactive artworks — pieces that run as code rather than sit as images. The premise is simple and a little stubborn: a generative artwork is not a picture, it is a small machine with knobs. Most of the web flattens it back into a JPEG. HIRO refused to. On HIRO you encounter the artwork live, you turn its knobs, and you leave with a version that is yours.

The idea came out of the DIY, on-demand design moment around 2020-2021 — Blush and EthBlock.art were the early reference points. The democratization of digital design had reached the point where people no longer wanted finished artifacts handed to them; they wanted a say in shaping the things they use. HIRO took that seriously for the corner of the world I cared about most: creative coders, whose work is interactive by nature and yet almost always gets published as a flat screenshot.

The hard part was technical, and it was the part I found most interesting. To let a buyer customize an artwork in real time, the artwork has to tell the platform which of its internal variables are safe to expose. So HIRO shipped a small SDK. With **Hiro Core**, an artist wraps their sketch and declares the parameters — a color, a speed, a count — that the editor is allowed to touch. With the **Hiro CLI** (`hiro push`), they publish it. HIRO then auto-generates an editor panel from those declarations: no bespoke UI per artwork, the controls fall out of what the artist chose to reveal. Buyers tweak, preview live, and export a unique rendition — free or paid, the artist decides.

The design carried as much weight as the engineering. Working with Anne Lee (formerly design lead at Framer), we spent four iterations on a single tension: the interface had to feel structurally like a tech product — an app store, a portfolio — while never competing with the art it framed. The rule we kept coming back to was *the UI should not override the artworks*. The recurring diagonal-X motif and the restrained component kit came out of that constraint, across roughly 80 screens and three breakpoints.

I'll write this honestly: HIRO did not get the launch it deserved. What started as a 2022 release slipped for years. Part of it was scope — a real-time editor for arbitrary generative code is genuinely hard, and for long stretches I was the only one building it. Part of it was that I was carrying a full-time job alongside it, and the budget for outside help thinned as the timeline stretched. It reached a fully implemented, tested, staged platform with the SDK open-sourced and a handful of artists onboarded — and then never quite crossed the line into a real public launch.

I keep it in the portfolio anyway, because the thesis still holds and still runs underneath everything I do now: software you are *inside* of, that adapts to the person using it, beats software you only look at. HIRO was an early, imperfect, multi-year attempt to build exactly that.