JSON to Base64: Encode JSON Payloads

How to encode JSON to Base64

1. Paste the JSON into Input · text, or press Sample to load an example object.
2. Copy the Base64 from Output · Base64. The size figures show how much longer it got.
3. For URLs, cookies or token-like values, tick URL-safe (- _) and No padding (=).
4. Download saves the result as base64.txt.

The sample on this page, {"id":42,"name":"base64hub","tags":["json","encode"],"active":true}, is 67 bytes and encodes to eyJpZCI6NDIsIm5hbWUiOiJiYXNlNjRodWIiLCJ0YWdzIjpbImpzb24iLCJlbmNvZGUiXSwiYWN0aXZlIjp0cnVlfQ==. To check a result, press Swap, or use Base64 to JSON, which also pretty-prints.

Whitespace and formatting change the Base64

Base64 encodes the exact characters of the JSON text, so two documents that mean the same thing produce different Base64 if a single space differs:

When the Base64 is compared, cached or signed, produce the JSON with one serializer and fixed settings, and encode that exact string. Key order is another trap: most serializers keep insertion order, so {"a":1,"b":2} and {"b":2,"a":1} encode differently.

Python and JavaScript disagree by default

The object {"name": "Zoë", "n": 1} serialized by each language’s defaults:

With URL-safe and no padding ticked, the JavaScript version becomes eyJuYW1lIjoiWm_DqyIsIm4iOjF9.

Where Base64 JSON is used

• HTTP headers. Custom headers carrying context, such as tracing or user metadata, often hold Base64 JSON, since raw JSON quotes and non-ASCII characters are unsafe in header values. Many servers reject header lines longer than about 8 KB, so keep these payloads small.
• URL parameters. OAuth state values, redirect targets and share links sometimes carry a small JSON object as Base64URL.
• Environment variables and secrets. A whole JSON config fits in one variable without escaping, for example a service account key in a CI system.
• Token formats. JWT header and payload are Base64URL-encoded JSON without padding. To build a signed token, use the JWT encoder; this page only encodes.

Base64 adds a third to the size, and the JSON’s own quotes and keys are often larger than the values. Base64 size overhead shows when compressing first pays off.

JSON to Base64 in code

const obj = { name: "Zoë", n: 1 };
const bytes = new TextEncoder().encode(JSON.stringify(obj));
btoa(String.fromCharCode(...bytes));                     // "eyJuYW1lIjoiWm/DqyIsIm4iOjF9"
Buffer.from(JSON.stringify(obj)).toString("base64url"); // "eyJuYW1lIjoiWm_DqyIsIm4iOjF9" (Node.js)

Without the TextEncoder step, btoa(JSON.stringify(obj)) encodes ë as a single Latin-1 byte (eyJuYW1lIjoiWm/rIiwibiI6MX0=), which UTF-8 decoders cannot read, and it throws on anything beyond Latin-1, such as an emoji.

import base64, json

obj = {"name": "Zoë", "n": 1}
text = json.dumps(obj, separators=(",", ":"), ensure_ascii=False)
base64.b64encode(text.encode("utf-8")).decode()   # 'eyJuYW1lIjoiWm/DqyIsIm4iOjF9'

More options on the JavaScript, Node.js and Python pages.

FAQ