Case Study
Pastry Corner Cafe & Lounge
A production restaurant site for a Chantilly, Virginia bakery & lounge, built end-to-end with seven MiOffice tools.
Scope
Client
Pastry Corner Cafe & Lounge
Location
Chantilly, Fairfax County, Virginia
Site
pastrycornercafe.com
Built by
JSVV SOLS LLC
Menu items modeled
292 rows + addon groups
Photo assets processed
53 raw → AVIF + WebP
Festival landings
Diwali, Eid, Christmas, Thanksgiving
Voice search
Speakable JSON-LD readouts
What we built
A restaurant marketing site with a structured menu, festival landing pages, and structured data for voice-search readouts. The site ships AVIF images with WebP fallbacks, Speakable schema for Google Assistant / Siri / Alexa, and 292 menu rows derived directly from the cafe's existing PDF menu — no copy-paste, no re-entry.
The work was done entirely in-browser using MiOffice. No image-editing desktop app, no separate OCR tool, no licensed stock photography. The deliverables for each surface — hero imagery, menu cards, OG images, Speakable readouts — came out of the seven tools below.
Tools applied during the build
- AI Image Generator
Festival hero artwork (Diwali, Eid, Christmas, Thanksgiving landings) — brand-palette-matched generations, no licensed stock.
- Image Compressor
53 raw bakery photos compressed to AVIF + WebP fallback — ~70% weight reduction, unlocking LCP wins.
- HEIC to JPG
iPhone-shot dish photos converted to web-compatible JPG before the AVIF pipeline.
- AI Document Summarizer
Legacy menu PDFs and catering brochures compressed into structured prose for /llms-full.txt content depth.
- PDF to Text
Existing menu PDF extracted to seed 292 menu item rows, addon groups, and the price ladder.
- Image Resize
Multi-size generation: OG images (1200×630), menu cards (800×600), hero strips (1920×1080) — single source asset.
- AI Voice Generator
Reference recordings for Speakable JSON-LD readouts — what Google Assistant, Siri, and Alexa say for voice queries.
Outcome
The site went live with full menu, four festival landing pages, and voice-search structured data on day one. Hero image weight came down ~70% versus the pre-AVIF baseline — a Core Web Vitals LCP win without compromising photo quality. Menu data flows from a single PDF source through PDF-to-Text and AI Document Summarizer into the structured menu rows shown on the site.
Visit the live site: pastrycornercafe.com.
Building a restaurant site?
The same MiOffice toolkit applies to any cafe, bakery, or restaurant. See the patterns and tools on the use-cases page.
AI tools for restaurant websites →