Readable text still needs its original boundaries

Encoded URLs and Query Strings: Read Each Layer

Encoding can make an ordinary URL look difficult to read. Decoding everything at once can make it easier to misunderstand. The useful approach is to keep the address intact, identify its components, and compare raw query text with a single readable decoding layer before drawing a conclusion.

Open the Link Inspector when you need a local view of one complete HTTP(S) address. It separates visible URL parts and reports findings, not a trust verdict. This guide explains how to read those parts without turning a display convenience into an unsupported claim about what a server will do.

Start with a complete address and respect the input limits

Paste the address itself, not surrounding quotation marks, an email label, or a piece of page markup. Select Inspect link to process it. Load example supplies practice text, and Clear removes the current work. The input requires the explicit HTTP(S):// form; a relative path, protocol-relative address, or bare hostname is not the requested input. Keeping that boundary clear matters because a missing scheme can make readers assume context that was never supplied. The inspector does not search for a website or choose a base address to complete an unfinished reference on your behalf.

All input limits apply together: no more than 8,192 Unicode code points, 32 KiB, or 200 query fields. A code-point limit and a byte limit measure different things, so an address is not accepted merely because its visible length looks modest. Controls, bidi control characters, backslashes, and internal whitespace are rejected. Outer ASCII spaces are trimmed, with that change reported as a finding. A pasted line break is not a harmless way to wrap a long address here. Obtain an intact copy rather than removing arbitrary characters from a real signed or private URL.

The initial result gives the ASCII and Unicode hostname, PSL-based registrable domain, scheme, effective port, and findings. It keeps private details hidden until you choose Reveal details. The domain boundary includes private Public Suffix List rules; it does not identify an owner. No API, fetch, DNS request, status lookup, reputation assessment, TLS test, availability probe, breach check, or ownership query runs against the supplied address. Successful parsing means that the text can be examined under these rules. It does not establish that a page exists, belongs to a particular organization, or should be opened.

Locate the path, query, and fragment before interpreting escapes

Reveal details toggles the normalized URL, path, ordered raw and once-decoded query pairs, and fragment. Read the component labels before reading individual values. In the invented address https://catalog.example/items?view=grid#specifications, items is in the path, view=grid is query text, and specifications is the fragment. These pieces do not become interchangeable because a decoded display makes them look similar. Keep the whole address available in the panel while you answer a narrow question about one component. This prevents a value copied from a detail row from being mistaken for a complete replacement URL.

A delimiter written as an escape inside a value is different from that delimiter separating components. A raw query value containing %26 can display an ampersand after decoding without creating a new query field in the original address. Similarly, an encoded question mark inside a value is not the outer URL query boundary. Read the raw structure first and the value second. If you replace encoded characters across the whole address before identifying boundaries, you can manufacture a different-looking structure and then accidentally review that invented version rather than the text someone actually supplied.

Normalization is another reason to keep the source context clear. The browser parser can produce a canonical representation, including a normalized hostname or removal of a default port from the serialized address. That is representation work, not a cleanup recommendation or a record of a network response. If a colleague asks whether the inspected text differs from their source, describe the relevant normalization instead of claiming the link changed destinations. Do not assume the normalized spelling is suitable for every signature-sensitive workflow. Keep the original authorized source available outside the inspector when exact source comparison is required.

Decode percent escapes once and leave malformed text visible

A percent escape represents a byte using a percent sign and two hexadecimal digits. The once-decoded query view makes encoded values easier to read without repeatedly unwrapping them. For example, %2F can represent a slash within a query value, while %252F becomes the visible text %2F after one pass. That remaining escape is useful evidence of another layer, not permission to decode again automatically. Stop at the layer the tool reports. A reader who keeps applying decoding until the result looks familiar can lose the distinction between the supplied value and a guessed interpretation.

Consider a filter value containing %2526. Its first readable layer contains %26, not a new outer query separator. Replacing that again with an ampersand and splitting the result would invent fields that were not separate in the supplied query. This is especially important when a value carries another address or a piece of structured text. Work from the displayed raw pair and explain the one-pass result in words. If another system intentionally uses multiple encoding layers, ask its author to describe that format rather than assuming every remaining percent sign is an error to remove.

Malformed encoding is reported and the raw text is retained. Examples worth questioning include a lone percent sign, %Q1, or an incomplete byte sequence. A failure to decode a value cleanly is not a reason to replace missing characters from memory. It can reflect a broken copy, an incorrectly generated link, or data whose intended format you have not been given. Use a sanitized example when asking the author about it. Do not claim the destination rejects the entire URL just because the inspector reports an encoding finding; no destination parser or server response was tested. When requesting a correction, identify the affected field and preserve the problematic raw sequence in a nonsecret example. Asking whether the producer intended a literal percent sign is more actionable than asking someone to fix a vaguely described strange-looking link.

Keep the raw spelling beside the readable value

Inspect one encoding layer at a time

Reveal ordered query details to compare percent escapes and plus signs without rewriting the supplied address.

Treat plus-to-space conversion as a stated convention

The readable query view applies the application/x-www-form-urlencoded convention in which a literal plus is read as a space. Under that convention, note=team+meeting displays the value team meeting, and note=team%20meeting also displays a space. The raw spellings remain different. This is not a universal promise about every server, framework, or application that may receive the address. When a support question concerns an exact plus character, give the raw spelling and identify the convention used by the readable view. A screenshot of only the decoded words can hide the distinction that matters.

An encoded plus and a literal plus also need separate attention. The raw value A%2BB displays A+B after the single form-decoding pass, while A+B displays A B under the same convention. A plus produced from %2B stays a plus during that same pass because the plus-to-space step has already happened. Repeating the complete form-decoding pass would instead turn the new plus into a space. This is a good reason to avoid repeatedly decoding, editing a decoded value and pasting it back into a complete URL without a proper construction step. You could change a literal plus into a space representation while believing that you merely made the address more readable.

Do not apply the query convention to every component. A plus in a path or fragment is not automatically a space merely because the query display uses form decoding. Likewise, a fragment that looks like a query may be data for the destination application rather than part of the outer query list. Discuss it under its actual component label. If the value contains a candidate embedded HTTP(S) URL in a named query field, the inspector examines that candidate locally only. It does not establish a redirect, follow a real flow, or expand a short link hidden behind another address.

Read query pairs as an ordered list, not a simple dictionary

The ordered list matters when a name occurs more than once. In tag=red&tag=blue, both occurrences are part of the supplied query. Do not collapse them to one value when describing the result. A receiving application might support a list or apply another rule, but the inspector cannot tell you which rule governs that destination. Reordering the fields for tidiness can also obscure a useful clue when comparing two supplied links. Keep the original order during review, especially when a developer needs to reproduce the exact structure rather than a simplified summary.

An equals sign separates a name from its value at the first occurrence in a field. For text=a=b, the value includes a=b; the second equals sign does not create another field. Distinguish empty assignments such as mode= from bare keys such as mode, and from no mode field at all. Also read an empty name carefully if the supplied query contains =value. These shapes can arise from templates or deliberate application conventions. Their presence can be reported without inventing the application meaning. A precise structural description is often more useful than calling the whole query broken.

Names can contain encoding too, so compare their raw and readable spellings before deciding whether they are truly different. A query that looks unique at a glance can still contain names that coincide in the decoded view. If a campaign template produced ambiguous fields, repair the template at its source with its author rather than deleting whichever row looks redundant. For a separate measured-link construction task, use the UTM Campaign Link Builder. For intentional cleanup, consider the URL Cleaner only after deciding what must remain. Neither task follows automatically from finding an encoded name.

Copy the intended representation and close the private work

Before copying, decide whether you need the complete inspected address or only its host. Copy inspected URL and Copy hostname are separate explicit actions. A readable query row is explanatory text, not a reconstructed full address. Inspected user URLs are not active anchors, so examining an escape does not require opening the destination. Keep private details hidden when you do not need them, but remember that the pasted input itself can still be sensitive. Secret material may appear in userinfo, path segments, query values, and fragments; a query-only review does not remove those other exposure points.

Local inspection takes place in an opaque-origin sandbox with no-store responses and connect-src none. That describes the inspection boundary, not all traffic associated with the surrounding site. Normal WordPress page requests are separate. The tool does not submit the pasted address for a remote diagnosis or retrieve its destination. Browser extensions, a hostile parent that replaces the interface, and device compromise are outside this protection. A decoded hostname or a clean-looking value does not change those limits. Treat findings as reading aids and do not convert them into a safety score or a promise about the page.

Copy before moving to another tab if the result is needed for an authorized handoff. The input and derived display clear when hidden and on pagehide, pageshow, or freeze; Clear also resets them directly. Returning to the page is not a way to restore the prior inspection. There is no history, saved project, download, PDF, or storage feature for later retrieval. Clearing the panel does not erase the clipboard or its history. After finishing, handle any copied full address according to its sensitivity, and share a sanitized field example when that is enough to explain the encoding problem.

Questions about Encoded URLs and Query Strings

Why does %252F still look encoded after the readable view?

The view decodes one layer. The percent escape %25 becomes a percent sign, leaving %2F as visible text. Another automatic pass would answer a different question about the value. Keep both the raw spelling and the one-pass result when asking the link author whether multiple layers were intended.

Does a plus sign always mean a space in a URL?

No. The query display uses the form-encoding convention, where a literal plus maps to space. That convention does not automatically apply to paths or fragments and does not prove how every server reads a query. An encoded plus, %2B, becomes a literal plus in the one-pass view.

Can I rebuild a working URL by copying all decoded values?

Do not treat the readable display as a reconstruction recipe. Decoded delimiters can become structural characters if pasted into a new address without suitable encoding. Use Copy inspected URL for an intentional full-address copy, or ask the author to produce a corrected link when its encoding is uncertain.

Continue with another link-inspection guide

Inspect a supplied address in the Link Inspector or choose a related task below. Explore more practical utilities in all free tools.

Preserve the distinction that matters

Read the URL without visiting it

Compare raw fields, note malformed encoding, and copy only the representation needed for your next authorized step.