Product search

Lasso Search lets you find products using natural language. Describe what you’re looking for and Lasso returns structured, typed product data matching your schema.

How it works

Describe what you need — Write a natural language query like “Sony noise cancelling headphones under $200”.
Lasso searches the web — The worker retrieves web search results that match your query.
Results are structured — Each product is extracted into your schema with typed fields, source URLs, and confidence scores.

Schema options

You control the output shape with three options:

Use an existing schema — Pass a schema_id from a schema you’ve already created via POST /schemas.
Define columns inline — Pass a columns array directly in the request.
Use the default — Omit both and Lasso uses a standard product schema (name, brand, price, description, category, image_url, url, features).

Sync vs async

By default, search is synchronous — you send a request and get results back in the same call. Pass a webhook_url to switch to async mode: Lasso returns 202 immediately and delivers results to that URL when ready. max_results defaults to 7 and accepts values from 1 through 7.

Example

const results = await client.search({
  query: "organic protein powder chocolate flavor",
  columns: [
    { key: "name", label: "Product Name", type: "text" },
    { key: "brand", label: "Brand", type: "text" },
    { key: "price", label: "Price", type: "number" },
    { key: "protein_per_serving", label: "Protein (g)", type: "number" },
  ],
  max_results: 7,
});

console.log(`Found ${results.total_results} products`);
for (const p of results.results) {
  console.log(`${p.data.name} by ${p.data.brand} — $${p.data.price}`);
}

results = client.search(
    query="organic protein powder chocolate flavor",
    columns=[
        {"key": "name", "label": "Product Name", "type": "text"},
        {"key": "brand", "label": "Brand", "type": "text"},
        {"key": "price", "label": "Price", "type": "number"},
        {"key": "protein_per_serving", "label": "Protein (g)", "type": "number"},
    ],
    max_results=7,
)

print(f"Found {results['total_results']} products")
for p in results["results"]:
    print(f"{p['data']['name']} by {p['data']['brand']} — ${p['data']['price']}")

Credits

Each search costs 5 credits, regardless of how many results are returned. The legacy model input does not select a provider. The service uses its configured Gemini provider. If the account lacks credits, the synchronous endpoint currently returns HTTP 200 with an error field and no results.

Next steps

Search API reference — Full parameter and response documentation.
Enrich — Take search results and enrich them with additional data.
Schemas — Create reusable schemas for consistent output.

Getting started

TypeScript

Python

Learn

Product search

How it works

Schema options

Sync vs async

Example

Credits

Next steps

​How it works

​Schema options

​Sync vs async

​Example

​Credits

​Next steps

How it works

Schema options

Sync vs async

Example

Credits

Next steps