Escape and unescape JSON strings: quotes, backslashes and line breaks

Here's what the tool gives for a short sample. It holds some quotes, a line break, a file path and a tab, and Escape mode turns the whole thing into one line.

He said "hi"
C:\temp\new	done

"He said \"hi\"\nC:\\temp\\new\tdone"

Read the escaped line one escape at a time. Each quote in "hi" becomes \". The line break becomes \n, which is a backslash and the letter n. The tab becomes \t. The file path has two backslashes, and each one doubles, so C:\temp\new is written C:\\temp\\new. The letters after those backslashes are plain text, so the \\t in the output is an escaped backslash followed by a plain t, not a tab.

In JavaScript, JSON.stringify of a string value gives this same quoted, escaped text. The rules are in RFC 8259 section 7, and they say:

• The quotation mark becomes \", and the backslash becomes \\. Both must be escaped.
• Control characters from U+0000 to U+001F must be escaped too. Five have short forms: \b for backspace, \f for form feed, \n for line feed, \r for carriage return and \t for tab. Any other control character uses \u and four hex digits, like \u0001.
• When you escape a character above U+FFFF, it becomes a surrogate pair of two \u escapes. U+1D11E becomes \uD834\uDD1E.
• A forward slash can be written \/, but RFC 8259 never requires it, and the tool leaves / alone.
• Other characters, like é and 😀, can stay as they are. RFC 8259 section 8.1 requires UTF-8 for JSON text exchanged outside a closed ecosystem, so no escape is needed.

Unescape works the other way. It takes the string with or without its outer quotes, so "{\"a\":[1,2]}" and {\"a\":[1,2]} both give {"a":[1,2]}. When the result looks like JSON, the Format it button formats it.

{"payload":"{\"a\":[1,2]}"}

A JSON document can hold another one as a string, which is the common reason to unescape. Here the payload value is a string, and its text is {"a":[1,2]}. A parser returns that text, not an object, so unescape the value with its quotes to get the inner document. Watch for four mistakes, and start with the first if your text came from a log or an API response, since it may already hold escapes:

• Double escaping. Escape doubles every backslash, so text that's already escaped, like a line from a log, gets escaped again. If the output shows \\\" where you expected \", the text went through Escape twice. Unescape text that already has escapes instead of escaping it.
• A raw line break. A control character can't sit unescaped inside a string, so a string that spans two lines fails. Unescape reports a "Bad control character" error and says a line break has to be written as \n. Replace the break with a backslash and an n, or run the text through Escape, which writes the \n for you.
• JavaScript escapes. \x41 and \' are valid in a JavaScript string literal, but JSON has neither. Unescape rejects both, because it runs JSON.parse. Write \u0041 for A, and drop the backslash before a single quote, since ' needs no escape in a JSON string.
• Python's default. json.dumps("é😀/") gives "\u00e9\ud83d\ude00/". The tool gives "é😀/" for the same input, and both mean the same text. The Python json docs say: "If True (the default), the output is guaranteed to have all incoming non-ASCII and non-printable characters escaped." Pass ensure_ascii=False to keep the characters.