Java escape and unescape

What a Java escape does, character by character

Javac reads a string literal from its opening double quote to the next one that is not preceded by a backslash, and it gives up with “unclosed string literal” the moment it meets a line break. Any text that holds quotes, backslashes or several lines therefore has to be rewritten before it can live in a .java file. A Java escape replaces each of those characters with a short backslash sequence the compiler turns back into the original: \", \\, \n, \r and \t. Paste the result between quotes, compile, and the String in memory holds exactly the characters you started with.

A single quote needs nothing inside a string literal, so it is left alone, and so are accented letters and emoji. Control characters without a letter of their own, such as NUL, come out as \u0000, which means the same thing whatever follows it. Octal \0 would not: "\01" is a single U+0001, not a NUL and a digit.

The unicode escape trap that javac sets before it reads a string

Java handles \uXXXX differently from every other escape. The compiler substitutes those sequences across the whole source file before it looks for strings, comments or keywords. So \u000a is not an escaped newline at all: it turns into a real line feed in the middle of the literal and the build fails. The same happens to \u000d, while \u0022 closes the string early and \u005c becomes a backslash that pairs with the next character. This page never writes those four characters in that form, and when you unescape it flags them as errors rather than decoding them into something the compiler would never accept.

The reverse problem matters too. If your text contains the six characters \u0041 as documentation or test data, the backslash is doubled so the compiler sees an ordinary backslash followed by letters. Several u characters in a row, as in \uuuu0041, are legal Java and are read back correctly.

String literal or text block

Text blocks arrived in Java 15 and look like the easy answer for multi-line text, but the compiler edits their contents in three passes. It converts every line ending to a line feed, removes the indentation that all lines share plus any spaces at the end of each line, and only then interprets escapes. Paste a paragraph with a trailing space or a Windows carriage return and the String you get back is quietly different.

Choosing Text block produces the complete literal. Line breaks stay real, a space at the end of a line becomes \s, carriage returns become \r, control characters use three-digit octal escapes that survive the stripping pass, and every third quote mark in a row is escaped so the block cannot close early. The closing delimiter sits in the first column, which tells javac there is no incidental indentation to remove. Indent the whole block together in your editor and the value stays the same.

Reading escaped Java back into text

Unescape follows the Java Language Specification: the short escapes including \s, octal values from \0 up to \377, and unicode escapes, in the order the compiler applies them. In text block mode it also strips indentation and trailing space exactly as javac does and honors a backslash at the end of a line, which joins two lines. What Java rejects is kept as written and listed with its position: \x41, \u{41}, \v, \q, a raw double quote or line break inside a string literal. Unlike JavaScript, Java never quietly drops the backslash of an unknown escape, so neither does this reader.

Carriage returns survive a file, not a paste

Whatever you paste into the box above has already passed through the browser’s form handling, which rewrites each CR LF pair as a single LF. That is standard browser behavior and no page script can undo it; the counter under the field shows the endings that arrived. To keep Windows line endings, open the file with the button instead. The bytes are decoded directly, each carriage return is written as \r, and the compiled String contains CR LF again.

Why C# and .NET strings are not handled here

Escaping tools often lump Java and .NET together, and that pairing hides real differences. A C# \x escape swallows up to four hexadecimal digits, so "\x41BC" is one character rather than an A followed by BC. C# also has \U with eight digits, \a and \v, but no octal escapes and no compile-time unicode substitution. Output escaped for one language can compile in the other and mean something else, so only Java is supported, and every Java behavior described on this page was checked by compiling test strings with javac 21.

Related escapes and formatters

When the string travels in an API payload rather than source code, escape it as JSON, which has no octal and no text blocks. For a Spring or Maven configuration file, XML escaping uses entities such as &, and a template that renders into a web page needs HTML escaping. To inspect a JSON or XML value before embedding it, run it through the JSON formatter or the XML formatter, and test a pattern for Pattern.compile in the regex tester before doubling its backslashes here.

Private by construction

The work is string processing in this browser tab. Nothing you enter is sent anywhere, logged, or written into the page address, and nothing is compiled or run. Connection strings, tokens and proprietary source are typical things to escape, so feel free to watch the network tab in your developer tools while you use it.