JSON Unescape — Decode Escaped JSON Strings Online

Paste an escaped JSON string and get the readable text back. Decodes \n, \", \\ and \uXXXX one layer at a time, handles double-escaped input, and keeps literal backslashes in Windows paths intact.

Escaped JSON turns up whenever a string has been through a serializer: log files, API responses, database columns, webhook payloads, CI output. This tool reverses that, entirely in your browser, so nothing is uploaded. If you need the opposite direction, use JSON Escape.

Updated September 2026 — dedicated unescape tool, split out from the combined escape page.

What is JSON unescaping?

JSON unescaping converts backslash escape sequences inside a JSON string back into the characters they represent. \n becomes a real line break, \" becomes a double quote, \\ becomes one literal backslash, and \u263A becomes ☺. It is the exact reverse of escaping.

Unescaping is not the same as deleting backslashes. A backslash disappears only when it is part of a valid escape sequence; one that stands for itself survives as a single backslash. That is why C:\\temp unescapes to C:\temp and stops there, and why find-and-replace on \ quietly corrupts file paths and regular expressions.

One pass decodes exactly one layer. A string that went through two serializers needs two passes — see double-escaped strings below.

JSON unescape reference table

What each escape sequence decodes back to:

Anything that is not a backslash sequence passes through unchanged.

How to unescape a JSON string

1. Paste the escaped text into the left editor. Surrounding quotes are optional — the tool accepts either.
2. Click Unescape → Right. One layer is decoded.
3. Check the output. If backslashes remain and you know another layer exists, click Unescape again.
4. Copy or download the result.

Input without surrounding quotes

Input (the two characters \n, not a line break):

Line 1\nLine 2

Output:

Line 1
Line 2

Input with surrounding quotes

This tool accepts a quoted string and keeps its surrounding quotes in the displayed output. The decoded line break is real, so what you see is not a valid JSON string literal to paste straight back into a payload.

"Line 1\nLine 2"

Output (surrounding quotes retained):

"Line 1
Line 2"

Remove escape characters from JSON

“Remove the escape characters” almost always means unescape, not strip every backslash. Clicking Unescape → Right removes the backslashes that are part of a valid sequence and keeps the ones that represent a literal backslash.

Deleting every \ with find-and-replace looks identical on simple input and silently corrupts file paths, regular expressions, and any text where a backslash is real data. If backslashes are still visible after one pass, the string carried more than one layer of escaping — decode it one layer at a time rather than removing them by hand.

How to fix a double-escaped string

A string acquires another layer of escaping each time it passes through a serializer. Decode only the number of layers you know were added.

Double-escaped input:

Line 1\\nLine 2

After the first pass (still a literal backslash followed by n):

Line 1\nLine 2

After Unescape again (a real line break):

Line 1
Line 2

Know when to stop

Unescaping C:\\temp\\file.txt once produces C:\temp\file.txt. Stop there if that is the intended path. Another pass would read \t as a tab and \f as a form feed, destroying the path. The Unescape again button appearing does not prove another layer exists — it only means escape sequences are still present, and in a Windows path they are supposed to be.

When unescaped output is not valid JSON

Unescaping returns plain text, not necessarily valid JSON. Decoded output can contain real line breaks and unescaped quotes, both of which are illegal inside a JSON string literal.

If your goal is a parseable object rather than readable text, use JSON Parse, which decodes the string and parses the result in one step. To check whether what you have is valid, use JSON Validator; to repair it, JSON Repair.

Unescape JSON in code

For a one-off the tool above is faster. In a pipeline, use your language's own parser rather than a regular expression — hand-rolled unescaping gets \uXXXX and surrogate pairs wrong.

JavaScript / Node.js

const decoded = JSON.parse('"' + escaped + '"');

Wrapping in quotes turns the fragment into a JSON string literal so the parser handles it. This assumes escaped is properly escaped; if it contains a raw unescaped quote it will throw, which is the correct outcome.

Python

import json
decoded = json.loads('"' + escaped + '"')

Java (Jackson)

ObjectMapper mapper = new ObjectMapper();
String decoded = mapper.readValue("\"" + escaped + "\"", String.class);

Go

var decoded string
if err := json.Unmarshal([]byte(`"`+escaped+`"`), &decoded); err != nil {
    return err
}

PHP

$decoded = json_decode('"' . $escaped . '"');

C# (.NET)

using System.Text.Json;
var decoded = JsonSerializer.Deserialize<string>("\"" + escaped + "\"");

Command line (jq)

echo '"Line 1\nLine 2"' | jq -r .

-r is raw output, which prints the decoded text rather than re-quoting it.

Common issues and fixes

• Backslashes remain after unescaping. Either the string was escaped more than once, or those backslashes are literal data. Decode one layer at a time and stop when the output looks right.
• “Invalid escape sequence”. JSON only defines \" \\ \/ \b \f \n \r \t \uXXXX. Sequences such as \x41, \' or \a come from other languages and are not valid JSON.
• Unicode escapes did not decode. \uXXXX needs exactly four hex digits. Characters outside the Basic Multilingual Plane, including most emoji, arrive as surrogate pairs like \uD83D\uDE80 and only decode correctly when both halves are present.
• A Windows path came out mangled. You ran one pass too many. \t in C:\temp became a tab. Start over and stop a pass earlier.
• Output will not parse as JSON. Expected — see above. Unescaping produces text, not a document.
• Nothing changed. The input had no escape sequences, or it was already fully decoded.

Unescaping vs URL decoding vs HTML decoding

A value copied out of a browser network tab has often been through more than one of these. Work outwards in the order they were applied, checking the result after each step.

FAQs