Skip to main content

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 →