Google Crawls Raw and Encoded Bracket URLs as Separate Pages, Mueller Confirms

Google's John Mueller has clarified that raw square brackets and their percent-encoded equivalents in URLs are treated as entirely different addresses — a distinction that could silently split your site's SEO signals without any warning.

The announcement came via Bluesky on October 21, 2026, when Mueller addressed a question from Vanessa Fox, a former Google employee who developed Google's Webmaster Central. Her question was simple but consequential: does Google treat ?filter[]=123 and ?filter%5B%5D=123 as the same URL? The answer, Mueller confirmed, is no — and he admitted he may have told people otherwise in the past.

---

Why Brackets Become a Technical SEO Problem

The distinction traces back to RFC 3986, the internet standard governing URI syntax. Under Section 2.2 of that standard, reserved characters — including square brackets and plus signs — cannot be swapped for their percent-encoded equivalents without creating a technically different URL. Mueller confirmed this by running a JavaScript comparison between the two URL forms, which returned false.

"I would have assumed they were the same and probably gave people the wrong answer on this in the past," Mueller wrote on Bluesky.

The technical nuance runs deeper than most SEOs might expect. RFC 3986 actually restricts the use of raw brackets in query strings entirely. According to the standard, brackets are only permitted in the hostname portion of a URL — specifically for wrapping raw IPv6 addresses. That means a URL like ?filter[]=123 is technically invalid by internet standards even before the encoding question arises.

Mueller also clarified that letter case within percent-encoding does not create additional URL variants. The values %5B and %5b are treated as equivalent under RFC 3986's rules for hex characters. So while capitalisation within encoding is not a concern, the raw-versus-encoded divide remains a live issue.

Understanding how reserved characters interact with crawl behaviour sits at the heart of core technical SEO principles and best practices — and this bracket distinction is a precise example of where standards compliance directly affects indexing outcomes.

The Browser Behaviour That Quietly Introduces Inconsistency

Mueller noted a browser behaviour that can introduce URL inconsistency without any deliberate action from the site owner. When a user types a bracketed URL directly into a browser's address bar, the browser may automatically encode the brackets before sending the request. However, if a bracketed URL is called through JavaScript or other code, the raw brackets may be sent exactly as written.

This means a single site can transmit both URL forms to Google without ever intending to create two separate addresses. The inconsistency is introduced silently, at the point of request, and is easily missed during routine audits.

His recommendation on this point was direct: "Always encode it, never show []."

---

What Google Actually Does When It Finds Both Versions

Mueller was transparent about Google's process when a server returns identical content for both the raw and encoded forms of a bracketed URL. Google will crawl both versions and, if the content matches — potentially at the checksum level — it will deduplicate them and select one version to index.

The problem is that Google chooses which version survives. If your internal links point to one form while your canonical tags reference another and your sitemap lists a third variation, Google receives mixed signals. That ambiguity hands control of a critical indexing decision to the algorithm rather than keeping it with the site owner.

This guidance is already reflected in Google's official URL structure documentation, which states that characters defined as reserved by IETF STD 66 — the standard containing RFC 3986 — must be percent-encoded. You can review the RFC 3986 specification directly via the IETF for the precise language governing reserved characters and their encoding requirements. Mueller's Bluesky clarification brings public awareness in line with what Google's own documentation has long required.

Signal Fragmentation and Why It Matters

When Google deduplicates two versions of what is functionally the same page, the consolidated signals — links, authority, crawl priority — should ideally flow to the canonical version. But when the signals themselves are split across two URL forms, deduplication becomes less reliable. PageRank, link equity, and crawl budget can all be distributed across variants that should never have existed as separate addresses.

Google's Non-Canonical Cabinet video series, which explores URL and indexing edge cases, had not yet published its second episode on robots.txt as of October 20, 2026. The first episode addressed emoji and non-ASCII characters in URLs — a related but distinct situation, since brackets carry reserved-character status under the standard while emoji do not.

---

What Site Owners Should Do Now

For sites already using raw brackets throughout their URL structure, Mueller offered measured reassurance. He said he would not expect ranking problems from an existing setup and acknowledged that Fox cited a real-world example of a site using invalid bracket notation that indexes without apparent issue. However, he was equally clear that he "wouldn't set up a new site like this."

Auditing for Bracket Inconsistencies

The practical priority for existing sites is consistency. Mueller recommended paying close attention to:

• Internal links — ensure all references use one consistent URL format
• rel=canonical metadata — confirm canonicals point to the encoded form
• Sitemaps — verify only one URL variant is submitted
• Structured data — check that URL references within schema markup are consistent

Mixed signals across those touchpoints are where the real risk lives. Adopting sound foundational SEO principles for long-term performance means treating URL consistency not as an edge case but as a baseline requirement for every site configuration.

Third-Party Crawlers and Tool Compatibility

Mueller also raised a concern that extends beyond Google's own crawler. He noted that, based on issues Google encountered more than a decade ago, some third-party crawlers and fetchers may not apply the same precise logic when processing bracketed URLs. He stopped short of naming specific tools but suggested site owners test the crawlers that matter most to their workflow.

This is a meaningful caution. If your SEO auditing tool normalises bracket variants before reporting, you may never see the split in your data — even while Google processes them as two distinct pages. Verifying crawler behaviour directly, rather than assuming consistency, is the more reliable approach.

For broader context on how Google's suite of tools surfaces crawl and indexing data, it's worth familiarising yourself with the most effective Google tools for SEO and business growth — several of which can help identify URL inconsistencies before they compound into indexing problems.

The Clear Default for New Development

For any new site or feature build, Mueller's guidance is unambiguous. Default to percent-encoding brackets in all query string parameters from the outset. Retrofitting URL consistency across an established site is considerably more complex than building it in correctly from the beginning.

For SEOs and site owners, Mueller's clarification points to three clear actions: audit your URLs to determine whether raw brackets appear in query strings; align every signal Google receives — links, canonicals, sitemaps, and structured data — around one consistent URL format; and default to percent-encoding brackets in any new development work. Consistency is the variable site owners can control, and in this case, consistency is what determines which version of a URL Google decides to keep.