/en/tool/url-encoder/ works Encodes in the browser
Paste a URL or parameter value — the tool encodes it with percent-encoding and decodes it back, breaks the URL into parts, displays the parameters in a table, and lets you rebuild the URL. It supports legacy encodings, Cyrillic domains, and a list of URLs, one per line. Everything is processed in the browser; nothing is sent to the server.
Each part separately: on the left, as written in the URL; on the right, decoded. Empty parts are not shown.
| Part | As in URL | Decoded | Characters |
|---|
The middle column shows the value as it appears in the link; you can edit the right column, and the middle column is recalculated immediately. The completed link is assembled below.
| Key | Value in the address | Decoded value | Remove |
|---|
The original string with every character converted to percent encoding highlighted, plus a byte-by-byte breakdown of the unique characters.
| Character | Code | Bytes | In the address | Count |
|---|
Three ready-to-use snippets for the selected mode, with your string inside. Copy and run them: the result will match what you see above.
Choose what to encode. “Entire address” preserves the structure: the colon, slashes, question mark, ampersand, and hash remain address delimiters. “Parameter value” encodes them too. “Form data” uses the x-www-form-urlencoded format, where a space becomes a plus.
If an old link decoded into garbled text, switch the byte encoding. The tool automatically tests UTF-8, windows-1251, and KOI8-R and suggests the matching one with a button—click it to recalculate the result.
See the “Address by parts” section: the scheme, username, password, host, port, path, query, and fragment are shown separately in decoded form. The Cyrillic domain is also converted to punycode there—in the address, it appears as xn--, not as percent encoding.
Edit the parameters directly in the table and click “Rebuild address”. The values are encoded again using the selected rules, and the link is placed in the input field. Remove an unwanted parameter with the button, or add a new one.
Process a list of addresses in batch: enable “Each line separately”, paste a column from a spreadsheet, and download the result as a .txt file. Parameters can be exported separately as .csv.
Share the page link together with the parsed result: the mode and string are appended to the page address after the hash.
character → bytes in the selected encoding → %XX for each byte
It is the bytes—not the letters—that are encoded. That is why the same text produces different percent sequences in UTF-8 and a single-byte encoding, and one cannot be substituted for the other. The host follows a different rule: percent encoding does not work there; the domain name is converted to punycode.
The input data is the same as the default field values: you can calculate it mentally and confirm that the tool is telling the truth.
The string has 10 characters: eight Cyrillic letters and two spaces.
Each Cyrillic letter in UTF-8 uses two bytes, which means six characters in the form %D0%BA. Eight letters produce 8 × 6 = 48 characters.
Each space becomes %20—three characters. Two spaces produce 6.
Total: 48 + 6 = 54 characters.
In “form data” mode, each space would become a plus: 48 + 2 = 50 characters. In windows-1251, each letter uses one byte, so the entire string would shrink to 8 × 3 + 6 = 30 characters.
The result is %D0%BA%D0%BE%D1%84%D0%B5%20%D0%B8%20%D1%87%D0%B0%D0%B9—54 characters instead of 10; all 10 original characters are replaced.
The set of characters that remain untouched and what happens to spaces. When encoding the entire address, the colon, slash, question mark, ampersand, and hash retain their role as delimiters. “Parameter value” encodes them too—otherwise an address inserted into a parameter would break your own address. “Form data” uses the x-www-form-urlencoded format: a space becomes a plus, and only the asterisk, hyphen, period, and underscore remain untouched.
These are two different rules. In the address itself, a space is encoded as %20. Form data uses the application\/x-www-form-urlencoded format, where a space is a plus. This tool supports both: the plus enables “form data” mode, and the plus is interpreted as a space when decoding.
Switch the byte encoding to windows-1251 or KOI8-R. Before UTF-8 became standard, Russian addresses used single-byte encodings, so %E0%E1%E2 is not a fragment of UTF-8 but three windows-1251 letters. The tool tests all three encodings automatically and, when it finds a match, shows a button with a suggestion.
Because the domain name system does not understand percent encoding: it works with Latin letters, digits, and hyphens. Punycode was created for other characters, so “сайт.рф” becomes “xn--80aswg.xn--p1ai”. Percent encoding applies to the path, query string, and fragment, but not to the host. In the “Address by parts” section, punycode is converted in both directions.
It also encodes the exclamation mark, apostrophe, parentheses, and asterisk. Built-in browser functions leave them as they are because they follow an older address specification, while PHP's rawurlencode and Python's quote encode them. The difference is rare, but it matters when the receiving side checks a signature byte by byte—which is why browser and server results can differ.
A percent sign in an address always starts a pair of hexadecimal digits. If anything else follows it, the sequence is malformed: most often, the address was encoded twice or its ending was truncated during copying. The tool shows the position of such a sign.
Decode repeatedly until the string stops changing. If an address was encoded twice, the percent sign itself became %25, and %D0%B2 became %25D0%25B2—one pass cannot undo this. With the option enabled, the tool makes as many passes as needed and reports how many layers it removed.
No. The browser handles encoding, address parsing, punycode, legacy encodings, and file downloads. After loading, the page sends no requests, and the string only reaches your own tab's address bar so you can share the parsed result as a link.
Creates passwords of the required length and assesses how difficult they are to crack.
Open