A little more room to read.
Wren searches for pages you can open and read without paying or creating an account.
Wren is in beta
That word is doing real work here, so here is what it means in practice.
- The index is small, and you will find holes. Wikipedia is complete; the rest of the web is a growing collection of measured sites, not a sweep of everything. Some searches come back thin, and a few come back with nothing.
- Some measurements are wrong. Checks use sampled pages, and a site can change after a visit. Loading speed varies across connections and runs, so it does not affect grades or rankings.
- It runs on one machine in a home. Occasionally it will be down. Nothing is lost when it is; come back later.
- Private query counts help us improve search. We count query text without visitor identifiers, as described in our privacy policy. If you choose to report a search, we also save the query and any note you send so we can investigate what is missing. Tell us what’s missing.
Search brings together Wikipedia and a growing collection of independent, specialist and news sites. We measure access, speed, advertising, readability and accessibility. Sites with barriers are excluded or marked with limitations; checks can miss things, and websites change. A site that lets you read only after you say who you are, such as a professional investor, is labeled with that condition.
We have measured 7,444 hosts. A measured host is not a guarantee that every one of its pages is indexed. You can inspect the evidence through each result’s site details.
Search tips
"sea level"requires those words together, in that order. You can combine a phrase with other words.PHP site:php.netsearches one site. Use the site domain shown beside a result; its subdomains are included.PHP -site:wikipedia.orgexcludes a site. You can exclude up to eight sites.site:php.netbrowses indexed pages from that site.
Site filters use whole site domains, not URL paths or individual subdomains. Quotes match words; punctuation and capitalization are handled like an ordinary search. Searches with phrases or site filters keep those requirements and do not add broader related results.
For C# and C++, pages with the requested symbols in their titles or nearby
text are prioritized where found. These are distinct indexed terms: a match for C alone cannot
satisfy either symbol, including in quoted searches. Broader matches retain the requested symbol.
A version after a software name, such as rxjava 3 or python 3.12, lists pages
whose title or address names that version before pages naming another version. A page's own title
counts over its address. Numbers after other words, such as windows 11 or
route 66, or a count such as python 3 ways, are searched as ordinary words.
A license, permit or registration search that names a U.S. state or a country, such as
new york fishing license, lists pages naming that place with the license and the activity, or that
place's own agency for the activity, first; then pages naming the place with the activity or
a license; and pages naming only another state or country lower. Court cases keep their ordinary position. Bare Georgia and Washington need a
qualifier, such as washington state. Neither is a filter: every matching page stays
in the results and the count.
Know what a grade means
A grade describes the reading experience, not the truth or neutrality of an article. Missing measurements stay missing. Read how grading works.
Your search is yours
No trackers, no ads and no third-party resources on the search page. We do not sell your queries or clicks, and money does not buy a place in the results. Our ranking promise.
Meet Pip
Pip is our little wren: curious, quick, and always looking for somewhere worth landing. That upright tail is your way home whenever you search.