This product was not featured by Product Hunt yet. It will not be visible on their landing page and won't be ranked (cannot win product of the day regardless of upvotes).
Product upvotes vs the next 3
Waiting for data. Loading
Product comments vs the next 3
Waiting for data. Loading
Product upvote speed vs the next 3
Waiting for data. Loading
Product upvotes and comments
Waiting for data. Loading
Product vs the next 3
Loading
ai2 Design System
Your agent writes the UI. This is the system it reads.
Every design system is complete until you need a small outline button in a warning tone. Then you are writing custom CSS, and the system you adopted for consistency is what makes you inconsistent. ai2 makes that button a prop. 51 MIT components, each with a real variant x tone x size matrix on one OKLCH token layer, light and dark from one source. 307 styled variants and 8 presets on top. Built for the machine writing your UI too: machine-readable axes, llms.txt, MCP. Free core, no account.
Every design system is complete until you need a small outline button in a
warning tone.
Then you are writing custom CSS at midnight, and the system you adopted for
consistency has become the thing making you inconsistent.
I am Berkay. I designed products for 15 years, watched that happen on every
team I worked with, and built ai2 so that button is a prop.
WHAT THAT MEANS IN PRACTICE
Every component carries a real variant x tone x size matrix. Not three
variants and a shrug. Button alone is 5 x 6 x 10 = 300 combinations, all
generated from one OKLCH token layer, all working in light and dark from a
single source. You stop leaving the system.
THE PART I CARE ABOUT MOST
I work with coding agents every day, and they kept getting component APIs
wrong. Not because the components were bad, but because nothing in a normal
library is written for a machine to read. So I built the registry the other
way round: machine-readable axes on every item, an llms.txt index, an
llms-full.txt that returns every component's axes and dependencies in one
request, and MCP so your agent installs and themes without you pasting
commands. The human docs fall out of the same source.
If your agent writes most of your UI now, this is the difference between it
guessing and it knowing.
FREE, MIT, TODAY
npx shadcn@latest add @ai2/button
51 components. 307 styled variations if the plain set is too plain. 8
presets. An accent theme generator. An inspiration gallery of 67 real sites
broken down into their tokens. A live extractor: give it a URL, get its
design tokens back. No account, nothing time-limited, nothing crippled.
THE PAID SIDE, PLAINLY
Pro is $99 (going to $149) for one seat, Ultimate $249 (going to $399) for
up to ten. One-time, lifetime updates, no subscription. The core stays MIT
either way.
Shipping today in Pro: 14 pricing blocks installable through the CLI, a Pro
MCP registry your agent can authenticate against, and every Framer product
I publish (Halcyon, Helio, Pulsar).
Still in development, and marked that way on the pricing page instead of
sold as done: premium themes, the full theme generator, a Figma kit with
variables in both modes, and Vite / Base UI / Radix starters. Ultimate adds
the Framer design system with tokens synced to the code side.
Tailwind v4, radix-ui, React 19. v0.7.0, so it is early.
Source: https://github.com/ai2design/ai2...
I would rather have criticism than upvotes. Tell me the axis naming is
wrong (variant / tone / size), or that a matrix this wide cannot stay
maintainable. Those are the two things I am least sure about.
About ai2 Design System on Product Hunt
“Your agent writes the UI. This is the system it reads.”
ai2 Design System was submitted on Product Hunt and earned 2 upvotes and 1 comments, placing #107 on the daily leaderboard. Every design system is complete until you need a small outline button in a warning tone. Then you are writing custom CSS, and the system you adopted for consistency is what makes you inconsistent. ai2 makes that button a prop. 51 MIT components, each with a real variant x tone x size matrix on one OKLCH token layer, light and dark from one source. 307 styled variants and 8 presets on top. Built for the machine writing your UI too: machine-readable axes, llms.txt, MCP. Free core, no account.
On the analytics side, ai2 Design System competes within Design Tools, Open Source, Developer Tools and GitHub — topics that collectively have 889.3k followers on Product Hunt. The dashboard above tracks how ai2 Design System performed against the three products that launched closest to it on the same day.
Who hunted ai2 Design System?
ai2 Design System was hunted by Berkay Erdinc. A “hunter” on Product Hunt is the community member who submits a product to the platform — uploading the images, the link, and tagging the makers behind it. Hunters typically write the first comment explaining why a product is worth attention, and their followers are notified the moment they post. Around 79% of featured launches on Product Hunt are self-hunted by their makers, but a well-known hunter still acts as a signal of quality to the rest of the community. See the full all-time top hunters leaderboard to discover who is shaping the Product Hunt ecosystem.
For a complete overview of ai2 Design System including community comment highlights and product details, visit the product overview.
Berkay Erdinc