Structured data extraction API: a field schema in, records out

curl -X POST "https://api.searlo.tech/api/v1/extract" \
  -H "x-api-key: YOUR_API_KEY" \
  -H "content-type: application/json" \
  -d '{
    "urls": [
      "https://shop.example.com/p/trail-runner-2",
      "https://shop.example.com/p/road-racer",
      "https://shop.example.com/p/discontinued"
    ],
    "schema": {
      "name": "string",
      "brand": "string",
      "price": "number",
      "currency": "string",
      "availability": "string",
      "rating": "number",
      "reviews": "integer"
    }
  }'

import requests

resp = requests.post(
    "https://api.searlo.tech/api/v1/extract",
    headers={"x-api-key": "YOUR_API_KEY"},
    json={
        "urls": [
            "https://shop.example.com/p/trail-runner-2",
            "https://shop.example.com/p/road-racer",
        ],
        "schema": {
            "name": "string",
            "price": "number",
            "currency": "string",
            "availability": "string",
            "rating": "number",
        },
        "useLlm": False,  # only values the sites declared; no model inference
    },
    timeout=180,  # URLs in one call are fetched one after another
)
out = resp.json()
print("billed", out["credits"], "credits for", out["totalResults"], "URLs")

for row in out["data"]:
    if not row["success"]:
        print("not billed, retry later:", row["url"], row["error"])
        continue
    print(row["url"], row["data"])
    print("  sources:", row["_sources"])                 # json-ld / meta / text / llm
    print("  missing:", row.get("missingFields", []))    # never guessed, always null

const res = await fetch("https://api.searlo.tech/api/v1/extract", {
  method: "POST",
  headers: { "x-api-key": "YOUR_API_KEY", "content-type": "application/json" },
  body: JSON.stringify({
    urls: ["https://shop.example.com/p/trail-runner-2", "https://shop.example.com/p/road-racer"],
    schema: { name: "string", brand: "string", price: "number", currency: "string", availability: "string" },
  }),
});
const out = await res.json();
console.log("mode:", out.mode, "| model available:", out.llmAvailable, "| credits:", out.credits);

for (const row of out.data.filter((r) => r.success)) {
  // Flag values a model inferred, so a person can review them before they ship.
  const inferred = Object.entries(row._sources)
    .filter(([, source]) => source === "llm")
    .map(([field]) => field);
  console.log(row.url, row.data, inferred.length ? `review: ${inferred.join(", ")}` : "");
}

{
  "searchParameters": {
    "type": "extract",
    "urls": 3,
    "fields": ["name", "brand", "price", "currency", "availability", "rating", "reviews"]
  },
  "mode": "structured+llm",
  "llmAvailable": true,
  "totalResults": 3,
  "data": [
    {
      "url": "https://shop.example.com/p/trail-runner-2",
      "success": true,
      "data": {
        "name": "Trail Runner 2",
        "brand": "Example Outdoor",
        "price": 129,
        "currency": "USD",
        "availability": "InStock",
        "rating": 4.6,
        "reviews": 312
      },
      "_sources": {
        "name": "json-ld",
        "brand": "json-ld",
        "price": "json-ld",
        "currency": "json-ld",
        "availability": "json-ld",
        "rating": "json-ld",
        "reviews": "json-ld"
      },
      "via": "http",
      "llmUsed": false
    },
    {
      "url": "https://shop.example.com/p/road-racer",
      "success": true,
      "data": {
        "name": "Road Racer",
        "brand": "Example Outdoor",
        "price": 89.5,
        "currency": "USD",
        "availability": "In stock",
        "rating": null,
        "reviews": null
      },
      "_sources": {
        "name": "meta",
        "brand": "llm",
        "price": "text",
        "currency": "text",
        "availability": "llm"
      },
      "missingFields": ["rating", "reviews"],
      "via": "browser",
      "llmUsed": true
    },
    {
      "url": "https://shop.example.com/p/discontinued",
      "success": false,
      "error": "scrape failed",
      "data": null
    }
  ],
  "source": "searlo.tech(extract)",
  "credits": 2
}