Geoffy Data Schema and API Overview

Integration Model

Today, Geoffy is primarily installed as a Shopify app. It connects directly to the store's data and does not require a separate public API key. The schema described here explains how the system works internally and how other tools can understand Geoffy's output.

Input: Shopify Product Data

At a minimum, Geoffy uses the following input fields from Shopify:

• Product title
• Product description
• Vendor / brand
• Product type
• Tags
• Variants (size, colour, etc.)

Here is a simple input example:

{
  "title": "Navy Satin Midi Wrap Dress",
  "description": "A flattering midi wrap dress in navy satin, perfect for evening events.",
  "vendor": "Fashion 9",
  "productType": "Dress",
  "tags": ["midi", "wrap", "evening", "navy"]
}

Output 1: GEO Bundle (Broad, Mid, Ultra)

Geoffy generates a GEO Bundle for each product, containing three tiers of content tuned to different specificity levels:

• Broad – generic category-level intent.
• Mid – category + one explicit attribute.
• Ultra – highly specific, purchase-ready intent.

Simplified example:

{
  "broad": {
    "searchTerms": ["womens dresses", "evening dresses"],
    "productContext": "A womens evening dress suitable for formal occasions.",
    "targetAudience": "Shoppers looking for a general evening dress.",
    "specificityScore": 3
  },
  "mid": {
    "searchTerms": ["midi evening dresses", "wrap evening dresses"],
    "productContext": "A midi-length wrap dress for evening events.",
    "targetAudience": "Shoppers narrowing down to midi or wrap styles.",
    "specificityScore": 6
  },
  "ultra": {
    "searchTerms": ["navy satin midi wrap dress", "Fashion 9 navy midi wrap dress"],
    "productContext": "A navy satin midi wrap dress by Fashion 9 for formal occasions.",
    "targetAudience": "Shoppers ready to choose a specific dress matching these attributes.",
    "specificityScore": 9
  }
}

This object is designed for AI systems to understand the product at different layers of intent, rather than as a single block of copy.

Output 2: fullContent Object

In addition to the GEO Bundle, Geoffy can generate a fullContent object for use on landing pages and in structured data. A fullContent object may look like:

{
  "heroSummary": "A navy satin midi wrap dress designed for evening occasions, offering a flattering fit and elegant drape.",
  "featureBullets": [
    "Midi-length wrap design for a flattering silhouette",
    "Smooth satin fabric that catches the light beautifully",
    "Versatile navy tone suitable for a range of formal events"
  ],
  "reasoningBlock": "This dress suits shoppers looking for a refined evening option that feels special without being overly formal. The midi length and wrap shape are comfortable for longer wear, while the satin fabric adds a polished finish.",
  "styleAndFit": {
    "fit": "Wrap design with adjustable waist tie",
    "material": "Satin outer fabric",
    "care": null
  },
  "searchMatchExplanation": "This product matches your search for a navy satin midi wrap dress suitable for evening occasions, combining the requested colour, fabric, length and style.",
  "faq": [
    {
      "question": "Is the dress lined?",
      "answer": "The dress is lightly lined for comfort without adding bulk."
    },
    {
      "question": "Can I wear this to a wedding?",
      "answer": "Yes, the dress is suitable for weddings and other semi-formal evening events."
    }
  ]
}

Fields may be null where source data does not provide enough information. Geoffy does not guess or hallucinate attributes.

Data Access for AI and Tools

Today, GEO content is primarily exposed via Shopify App Proxy landing pages and structured data (for example, JSON-LD). External tools can discover and understand GEO Bundles by crawling the relevant sitemaps and parsing the JSON structures on those pages.

Over time, Geoffy is designed to support more direct, authenticated integrations for AI systems and partner tools that need store-level GEO data.

Compliance and Limitations

Geoffy follows strict compliance rules:

• No hidden AI-only content.
• No attribute hallucination.
• Structured data always reflects visible product information.

For questions about deeper integrations, please contact us.