AI NewsGo-to-marketReported
Google counts raw and encoded brackets in a URL as different URLs
John Mueller said on Bluesky that a URL with raw brackets and the same URL with percent-encoded brackets count as two different URLs, so a site that mixes the two forms splits its signals across both and should pick one.
Why it mattersA site that uses both forms splits ranking signals across two URLs that look the same to a reader, which is a cheap, invisible way to lose indexing coverage until a canonical pass clears it up.
A URL that looks the same to a reader can be two URLs to Google. Search Engine Journal reports that John Mueller, Google's search advocate, said on Bluesky on Monday that raw square brackets and their percent-encoded equivalents count as different URLs, and that he had probably given people the wrong answer on this in the past.
The exchange was with Vanessa Fox, a former Google employee, who asked whether Google treats ?filter[]=123 and ?filter%5B%5D=123 as the same URL. Mueller said: "Raw and percent-encoded brackets in a URL count as different URLs." He pointed at RFC 3986, section 2.2, which treats brackets as reserved characters, so a reserved character and its percent-encoded form are not equivalent.
When Google de-duplicates them
Mueller said Google does find and merge the pair when the content on both URLs matches. "We'd crawl both versions, if we see the same content (maybe even checksum level), we'd de-dupe them and pick one." His recommendation is one line: "always encode it, never show []". Barry Schwartz at Search Engine Roundtable quotes Mueller noting that square brackets are "technically invalid" in a path, because the RFC only permits them in a hostname to wrap a raw IPv6 address.
The practical risk is a listing page whose filter links produce both forms through different code paths. A reader clicks a filter and gets a working page either way, and a site owner reading only a browser bar never sees that the sitemap and the internal links are pointing at a different URL from the one Googlebot found through a referral.
What to change
Search Engine Journal's write-up lists four actions. Use one form across every internal link, every canonical tag, every sitemap entry and every structured-data url field. Encode brackets using %5B and %5D. Audit internal links and meta tags for consistency. Do not panic about an existing site with raw brackets, but stop generating new URLs that way.
Mueller is clear that the fix is for new URLs, so a working site does not need a mass rewrite. A canonical tag that points at the encoded form, together with a sitemap entry that uses the same form, is enough to tell Google which version to keep.
For a team running a listing site or a search product, this is a small bug that is easy to miss and easy to fix. The templates that build filter URLs and the ones that build canonical tags often live in different parts of the codebase, and the two can drift without an obvious failure in production. Reading a sample of the generated URLs for one filter, then checking the canonical tag the same page emits, is a five-minute audit that answers whether the site is affected at all.
Source
Google Crawls Raw And Encoded Bracket URLs Separately, Search Engine Journal, by Matt G. Southern.
Also: Google Had Internal Issues With Square Brackets In URLs, Search Engine Roundtable, by Barry Schwartz. RFC background: RFC 3986, section 2.2 Reserved Characters.
This item was written by an AI system from the linked source. Reveneau is responsible for what it publishes.
Get AI News in your inbox
New developer tools, model and agent releases, and how teams are actually using them to release software. Short, and only when there is something worth reading.
