Manycrawl

Templates / Research

Google results

Scrape Google search results into a table. Each result becomes a row, and “Scrape more” picks up from the next page.

Source Research Desktop app

What it isSource

Sources put rows into a table. This one is the starting point of a list, not a column on an existing one.

InputWhat you give it

Input

A search — this one creates the rows rather than reading them.

OutputWhat comes back

  • Title, URL and snippet
  • Result type and position
Output example real run
PositionTypeTitleURLSnippet
1adThe best CRM for startupsattio.comThe best CRM for startups
2adCRM Tool For Startup — Smart Tools For Startupspipedrive.com/en/gettingstarted-crmCRM Tool For Startup — Smart Tools For Startups
5organicWhat's the best CRM for an early stage, bootstrapped…reddit.com/r/CRM/comments/1ncf59q/We're just 2 people right now who are growing fast averaging 60 leads a month…
6organicBest CRM For Startups: The All-In-One Guidesalesforce.com/eu/crm/startup-crm/Starter and Pro Suite are two Salesforce solutions that were specifically designed for startups…

Real output. "best crm for startups", page one. Ads and organic results are both captured and labelled, so you can keep or drop either.

How it worksThree things you decide

  1. Type the queryExactly as you would in Google — operators included. site:, inurl: and quoted phrases all work.
  2. Take the first pageEvery result becomes a row, labelled organic or ad, with its true position.
  3. Scrape more when you want depthPicks up from the next page rather than starting over, so nothing is fetched twice.

Limits — read this before you run it

Runs in a built-in browser, no tabs. A rotating proxy is recommended — search is where Google notices repetition. Each new result costs 1 action.

FAQQuestions people actually ask

Do search operators work?

Yes — site:, inurl:, quoted phrases and the rest, exactly as you would type them into Google.

How do I get past page one?

Scrape more picks up from the next page rather than starting again, so nothing is fetched twice.

Do I need a proxy?

It is recommended here. Search is where Google notices repetition fastest, so a rotating proxy makes long runs much steadier.

Are ads included?

Yes, labelled. Every row carries its type, so you can keep ads as a buying signal or filter them out.