TeleScope API

Semantic Address API

Query any product's semantic address, where it lands in AI meaning-space, verified across multiple models.

Endpoint

GET /api/public/address?store={domain}&product={handle-or-id}
GET /api/public/address?text={raw text}

Response 200 (application/json):
{
  "atlas_version": "v1.0",
  "product": {
    "title": string,
    "store": string,
    "id": string,
    "gtin": string | null,
    "sku": string | null,
    "url": string
  },
  "address": {
    "top_concepts": [
      { "concept": string, "category": string,
        "openai_similarity": number, "gemini_similarity": number }
    ],
    "cross_model_agreement": number,   // Pearson r across models
    "placement_score": number,         // 0-100
    "category_verdict": string,
    "flags": string[]                  // off_panel | thin_text | models_disagree
  }
}

Try it

curl "https://anchor-map-match.lovable.app/api/public/address?store=nitinkumarbooks.com&product=mastering-web3-and-crypto-for-investors-and-builders"

Live example (cached)

nitinkumarbooks.com · mastering-web3-and-crypto-for-investors-and-builders

{
  "atlas_version": "v1.0",
  "product": {
    "title": "Mastering Web3 and Crypto: For Investors and Builders",
    "store": "nitinkumarbooks.com",
    "id": null,
    "gtin": null,
    "sku": null,
    "url": "https://nitinkumarbooks.com/products/mastering-web3-and-crypto-for-investors-and-builders"
  },
  "address": {
    "top_concepts": [
      {
        "concept": "web3 blockchain book",
        "category": "Books & media",
        "openai_similarity": 0.71,
        "gemini_similarity": 0.69
      },
      {
        "concept": "cryptocurrency investing guide",
        "category": "Books & media",
        "openai_similarity": 0.68,
        "gemini_similarity": 0.66
      }
    ],
    "cross_model_agreement": 0.81,
    "placement_score": 72,
    "category_verdict": "Books & media",
    "flags": []
  },
  "_note": "Live scoring unavailable, showing cached example."
}

Use cases

Four ways teams wire one JSON endpoint into an existing pipeline.

AI shopping agents

Verify what a product IS before recommending it.

GET /api/public/address?store=acme.com&product=running-shoe-42
→ address.top_concepts[0].concept === "trail running shoes"
→ product.gtin, product.sku   // identity keys to hand off

Feed & catalog tools

Validate placement before publishing. Fail the CI check on mismatch.

# pre-publish check
addr=$(curl -s ".../address?store=$STORE&product=$H" | jq -r \
  '.address.category_verdict')
[ "$addr" = "$DECLARED_TYPE" ] || exit 1

Agencies & SEO apps

Attach cross-model placement evidence to client reports.

// before rewrite
before = await addr(store, handle)
// after rewrite
after  = await addr(store, handle)
delta  = after.placement_score - before.placement_score

Marketplaces & aggregators

Normalize third-party catalogs onto one category grid.

// each item lands on the same taxonomy
item.category_path
// e.g. "Apparel & Accessories > Shoes > Athletic Shoes"
// (Google + Shopify taxonomy, cross-model verified)

The shared map

Every audit computes a store's semantic address on one shared grid. As the map fills in, any agent can look up any store, one address system for AI commerce.

0 stores mapped · Panel v3.0.1