Skip to main content
The Search Playground runs a real search query and gives you back the raw results: titles, links, descriptions, thumbnails, exactly as a search engine returned them. Nothing is summarized, rewritten, or filtered by AI. It’s the fastest, cheapest way to get structured search data out of Spidra, and the right tool when you want to know what’s out there before deciding what to scrape.
This is deliberate. Search and Scrape solve different problems. Search answers “what results exist for this query?” Scrape answers “get me the content of this specific page.” If you already know the page you want, use Scrape directly. Search is for discovery.
1

Enter a query

Type what you’re looking for, the same way you would on any search engine. Natural questions and short keyword queries both work.
2

Choose your sources

Pick which kinds of results you want back: Web, News, Images, or Videos. You can select more than one. Each runs independently, so one source coming back empty doesn’t affect the others.
3

Set a country (optional)

Choosing a country does two things at once. It localizes the results the way a real user in that region would see them, and it routes the request through a proxy exit point in that region. Leave it on Worldwide for generic results.
4

Set a result limit

Control how many results come back per source, from 1 to 20. Lower limits resolve faster and cost less.

​
Understanding results

Results are grouped into tabs matching the sources you requested. Switching tabs never mixes data. The JSON view, the copy button, and the download button all reflect exactly the tab you’re looking at, not the whole response.
  • Results view. A readable card layout, with a Scrape page button on each web/news result that opens the Playground with that URL pre-filled, so you can go from “found it” to “scrape it” in one click.
  • JSON view. The raw response for the active tab, ready to paste into your own code.
  • Download. Results/JSON tabs export as CSV or JSON matching what’s on screen. The Images tab works differently on purpose: since the results are images, download bundles the actual image files into a ZIP instead of a CSV of URLs. Individual images can also be downloaded one at a time.

​
Features

  • Multiple sources in one call. Request web, news, images, and videos together instead of running separate queries.
  • Automatic retry. If a request gets blocked, Spidra automatically retries against an alternate result provider behind the scenes. You never see the retry or configure it, you just get a result.
  • Geo-targeted results. The country parameter controls both localization and the proxy region the request appears to come from, so results genuinely reflect what someone in that country would see.
  • No AI, no extra cost for it. Search never calls an LLM. Results are exactly what came back from the query, which keeps it fast and cheap compared to an AI-extraction job.
  • Search history. Every search is logged with its query, sources, and results, viewable and re-openable from your Logs.

​
Search vs. Scrape vs. Crawl

A common pattern: run a Search, review the results, then click Scrape page on the one that looks right. No need to copy and paste the URL yourself.

Submit a Search

Full API reference for the search endpoint

Search Logs

See your search history via the API

​
Tips

  • Start with just web. It’s the fastest source and covers most queries. Add news, images, or videos only when you actually need them.
  • Use Search before Scrape when you’re not sure of the URL. Searching first and scraping the right result afterward is usually faster than guessing a URL.
  • Set country when locale matters. Local news, region-specific pricing, and “near me” style queries all depend on where the request appears to come from.
  • Keep result limits modest. Most use cases don’t need the full 20. A smaller limit resolves faster.