Websites by topic
like: finds websites about a topic. Describe the topic in a few words,
in any language, or give a website about it, and the websites most like it come
first. Added to an ordinary query, it puts that query's results in order of
similarity instead of popularity rank.
like: is the topic of a website, not its code. Websites with the same
code as given ones are found by their cluster
signatures.
like:"words" - about a topic
Words in quotes describe the topic directly, in any language and alphabet. The websites do not have to use these exact words - the meaning is what counts.
like:"wedding photographer"
like:"Zahnarzt in Berlin"
like:"Tierarzt" site:de
like:example.com - about the same thing
Compares what the home page says, in any language: the same business, field or topic, wherever the website is and whatever it is built with.
Several examples - similar to all of them
Separate domains with commas, without spaces, or write like: once for
each example - that way domains and words in quotes can be mixed. A website has to
be like every example, not just one of them: the more examples, the closer the first
rows are to what they all share. like: still only orders the results - the total
stays the same. Up to 10 examples in one search.
like:jquery.com,getbootstrap.com
like:"recipes" like:"vegan"
With an ordinary query - its results, most similar first
Add like: to any query. The query decides which websites are in
the results; like: decides their order.
"wp-content/plugins/elementor" like:"yoga studio"
"static.hotjar.com" like:"online course"
"googletagmanager.com/gtm.js" like:"Zahnarzt" site:de
like:chefkoch.de site:de
An order, not a filter
Without like:, results come in order of popularity rank. With it, they
come in order of similarity, from the most similar down. With an ordinary query you get
the same websites and the same number of results as without like: -
only the order changes. Without a query, like: lists the websites your
plan opens, most similar first.
The top is broad: neighbouring topics rank high too. To narrow it, add a phrase the websites must contain.
Forms
- The operator can be written in any case:
LIKE:works too. - A domain can be written as it appears in the address bar:
https://,www.and the path are ignored. The home page of the website is what gets compared. - Words go in quotes:
like:"wedding photographer". Without quotes the word is read as a domain. - Inside quotes the operator is ordinary text:
"like:jquery.com"looks for that string in the source code.
Results
- Only home pages are compared, one result per website.
- The example itself is left out. Copies and mirrors of it stay - nothing is more similar.
- Equally similar websites are ordered by popularity.
- The example has to have its home page in the index, with enough text on it;
otherwise the search says so and returns nothing. Shops and landing pages often
have little text on the home page: when
like:example.comfinds nothing, say what the website is about in words -like:"handmade leather bags". - Without a query there is no code to show, so each result shows the title of its home page.
- In any search results, the (i) after a domain opens its page: the technologies found on it and the websites most like it.
Which websites are compared
like: compares the websites your plan opens in search results: the top
1,000,000 by popularity for a guest, the top 3,000,000 for a free account, and every
website on a paid plan. On a guest or free account the results say which websites
were compared. With an ordinary query, its results outside that range keep their
place after the compared ones, in order of popularity, as hidden rows.
- As many results as an ordinary search, and as many pages as your plan opens - the first one without signing in, 50 on a free account (see Pages of results) - or take them through the API up to your plan's limits.
- A search with
like:counts against your daily searches like any other search.