Google Search Light Workflows

Typical Light workflow

1. Submit POST /v1/google/search/light with query (or url) and options.
2. Read pages in the same response.
3. Check pages_returned and pages_stopped for multi-page requests.
4. Optionally pass result URLs into Extract, Batch, or Map (separate Scraper plan and balance).

Light is synchronous. You do not poll for completion.

Light vs general Search

Light and general Search use different endpoints and response contracts. Use the Light examples in these guides for Light requests.

Common patterns

Single-page JSON

Minimal query, default format: "json", one page billed when successful.

Multi-page in one call

Set pages (1–10). Same session and exit IP. Charge only returned pages. Inspect pages_stopped if fewer pages came back. Not with page. A request with both returns 422.

HTML or both

Use format: "html" or "both" when you need the raw Google page. JSON+HTML does not double the page charge.

AI Overview

Set ai_overview: true for page 1. If Google has no overview, results still return. When Google loads the overview after the page, we fetch it too, which can add a few seconds.

Location-sensitive search

Pass encoded uule only. Plain city names are not accepted in uule.

Hand off to Extract

Light returns organic URLs. Extracting page content uses the Extract product and its own balance.