ShopatLLMs
Gets Indian D2C brands recommended inside ChatGPT, Perplexity and Gemini.
Visit live site ↗
- Role
- Product, design & build
- Type
- AI commerce product
- Stack
- Astro · React · Tailwind · GSAP · Lenis · pnpm monorepo
- Live at
- shopatllms.com ↗
(01)The problem
Shoppers increasingly ask an AI assistant what to buy. Most Indian D2C brands have no llms.txt, thin product schema and no AI-readable feeds, so the assistant recommends someone else.
(02)Approach · key decisions
01
Lead with a free scan
A no-signup AI-visibility scan is the first thing on the page, so a founder sees their own gap in seconds before reading any pitch.
02
Explain GEO against SEO
A side-by-side table maps what founders already know (SEO) to the new game (GEO), so nothing needs a glossary.
03
Show the files, not the buzzwords
llms.txt, schema.json and ACP / UCP / Perplexity feeds appear as a real file tree, which makes an abstract service concrete.
04
A calmer, editorial redesign
Rebuilt after midday.ai: serif headlines, square corners, hairline sections, a dashboard mock in the hero and a vertical How-it-works beside a live phone demo. Copy now lives in one data file.
05
Tiered pricing by layer
Starter ‘get found’, Growth ‘get bought’ and Agency for portfolios. Each tier is one clear promise.
(03)What's inside
The build.
- Live ChatGPT-style demo card that shows a brand being recommended
- llms.txt setup, JSON-LD product schema, ACP / UCP / Perplexity product feeds
- Competitor tracking and real-time catalogue sync in the product story
- Planned ‘Buy now inside AI’ and WhatsApp buy flow
- Mega-menu navbar, results table, guarantee checklist and coverage marquee
Mobile · 390px
Desktop · 1440px
(04)Outcome
- ✓ Shipped
Live product site with free scan
- ✓ Shipped
Three-tier pricing: ₹1,199 · ₹4,499 · ₹29,999 / month
What I learned
Selling a new category is mostly translation. The page got clearer every time I swapped a term for a picture of the actual file.
Next project →
BOB Ads
