SIL
Shopping Interface Layer · for your AI agent
Open beta
Turn your Personal AI into an expert Shopping Agent
Shop like you code.
SIL is a shopping agent plugin for your personal AI host. Give it a spec, your last width, instep and flex, and it shops the stores that speak UCP and hands back a ready-to-buy link on the store's own site. It runs on OpenClaw today, in open beta: free, open source, and still changing.
Spec Driven Shopping
You're not a bad shopper. You're a human one.
Same goal: the right ski boot. Watch what each of you actually does.
searching like a human
- best ski boots for intermediate skiers
- Tabs open: 15Top 10 ski boots (2026), Reddit: size down or not?, Boot sizing charts, Flex index, explained
- size downstiffer is bettercomfort first
- add to cart… maybe?
- Slope 3. Shin bang. Day over.
shopping like you code
- get me new ski boots, for me
- Expert · Ski-Boot Fitter
- fitting model … readyyour foot … 27.5 · high instep · D-widthyour skiing … intermediate · all-mountain · warmth › weight
- matched on your last width, instep and flexAll-mountain boot, high-instep last· example pick· matched on instep and last width· checkout link on the merchant's site
- Happy skiing. All day.
The agent didn't search harder. It shopped from a spec.
You did the measuring
Your numbers are the input, not the search term
You know your last width, your instep, your chest. A search box has nowhere to put them, so it takes the size label instead, and the label is a lossy compression of the geometry you just measured. Your agent does not have that problem. It reasons over the measurements themselves.
It shops your last width, your instep, your flex, not the marketing.
How SIL works
Set up an expert shopper
Mint a specialist for a niche you care about: ski touring, road cycling, espresso. Or skip it and shop with a generalist; specialize anytime.
Shop from the spec, not the search box
Your agent takes your requirement to the stores that speak UCP, the Universal Commerce Protocol, and holds every candidate against it.
Compare and fit, with SDS
Spec Driven Shopping reasons over your specs, not the marketing, to find the genuine best fit and the true value.
Buy with confidence
A ready-to-checkout link on the merchant's own site, every pick backed by a why.
You do step one. Your agent does the rest.
The payoff
Transforms the shopping experience
Your Agent
Best bang for the buck
SDS reasons over what actually matters in the category, so it uncovers the true value of what you're buying, not the loudest listing.
Ideal fit, personalized
It optimizes the dimensions that produce real fit: your body, your taste, your level, so it's right the first time.
You know before it ships
The reasoning is on the table before you buy: what it matched, what it rejected, and why. No surprise waiting in the box.
It learns you, continuously
Every rejection sharpens it. Next month's expert knows you better than today's.
Ultimately, it's yours.
You shape it, you own it. Your specs and your experts live with you, carried across hosts, never a platform's profile of you. The more you shop, the more it becomes unmistakably yours.
openclaw
# install the plugin for OpenClaw
openclaw plugins install clawhub:@4gpts/sil# then two lines and you're shopping
> register me on SIL
> make me a shopping expert for ski touring
Not on OpenClaw?
We are working on SIL for Meta's Muse and for Hermes Agent next. Leave your email and we write once, when it runs there.
Get started
Install once. Then shop like you code.
OpenClaw today, in open beta. Your expert shops the UCP stores and hands back a ready-to-buy link on the merchant's own site.