package
Yet another Base64 transcoder in pure-JS
npm install js-base64GitHub-flavored package documentation.
Across all versions.
Download trend unavailable.
Runtime metadata.
Tagged topics.
Estimated unpacked size across 1 packages.
39.0 kB
Includes dependencies.
Yet another Base64 transcoder.
$ npm install --save js-base64Locally…
<script src="base64.js"></script>… or Directly from CDN. In which case you don't even need to install.
<script src="https://cdn.jsdelivr.net/npm/js-base64@3.9.6/base64.min.js"></script>This good old way loads Base64 in the global context (window). Though Base64.noConflict() is made available, you should consider using ES6 Module to avoid tainting window.
locally…
import { Base64 } from 'js-base64';// or if you prefer no Base64 namespace
import { encode, decode } from 'js-base64';or even remotely.
<script type="module">
// note jsdelivr.net does not automatically minify .mjs
import { Base64 } from 'https://cdn.jsdelivr.net/npm/js-base64@3.9.6/base64.mjs';
</script><script type="module">
// or if you prefer no Base64 namespace
import { encode, decode } from 'https://cdn.jsdelivr.net/npm/js-base64@3.9.6/base64.mjs';
</script>const {Base64} = require('js-base64');Unlike the case above, the global context is no longer modified.
You can also use esm to import instead of require.
require=require('esm')(module);
import {Base64} from 'js-base64';let latin = 'dankogai';
let utf8 = '小飼弾'
let u8s = new Uint8Array([100,97,110,107,111,103,97,105]);
Base64.encode(latin); // ZGFua29nYWk=
Base64.encode(latin, true); // ZGFua29nYWk skips padding
Base64.encodeURI(latin); // ZGFua29nYWk
Base64.btoa(latin); // ZGFua29nYWk=
Base64.btoa(utf8); // raises exception
Base64.fromUint8Array(u8s); // ZGFua29nYWk=
Base64.
Base64.decode( 'ZGFua29nYWk=');// dankogai
Base64.decode( 'ZGFua29nYWk'); // dankogai
Base64.atob( 'ZGFua29nYWk=');// dankogai
Base64.atob( '5bCP6aO85by+');// 'å°é£¼å¼¾' which is nonsense
Base64.toUint8Array('ZGFua29nYWk=');// u8s above
Base64.decode( '5bCP6aO85by+');// 小飼弾
// note .decodeURI() is unnecessary since it accepts both flavors
Base64.decode( '5bCP6aO85by-');// 小飼弾Base64.isValid(0); // false: 0 is not string
Base64.isValid(''); // true: a valid Base64-encoded empty byte
Base64.isValid('ZA=='); // true: a valid Base64-encoded 'd'
Base64.isValid('Z A='); // true: whitespaces are okay
Base64.isValid('ZA'); // true: padding ='s can be omitted
Base64.isValid('Z'); // false: a single character can't encode a byte
Base64.isValid('++'); // true: can be non URL-safe
Base64.isValid('--'); // true: or URL-safe
Base64.isValid('+-'); // false: can't mix bothThe default encoding uses the standard Base64 alphabet from RFC 4648, section 4. The URL-safe form in js-base64 uses the base64url alphabet from section 5 and omits padding.
| Form | Characters for values 62 and 63 | Padding | Encoding methods |
|---|---|---|---|
| Standard Base64 | + and / | Trailing = as needed | Base64.encode(text), Base64.fromUint8Array(bytes) |
| URL-safe Base64 | - and _ | Omitted | Base64.encodeURI(text), Base64.encode(text, true), Base64.fromUint8Array(bytes, true) |
For example, these bytes show both the alphabet and padding differences:
const bytes = new Uint8Array([251, 255]);
Base64.fromUint8Array(bytes); // '+/8='
Base64.fromUint8Array(bytes, true); // '-_8'
Base64.toUint8Array('+/8='); // Uint8Array [251, 255]
Base64.toUint8Array('-_8'); // Uint8Array [251, 255]Base64.decode() and Base64.toUint8Array() accept both alphabets, with or
without padding. Use decode() for UTF-8 text and toUint8Array() for bytes.
Encoded output has no line wrapping; MIME or PEM line wrapping is not added.
By default Base64 leaves built-in prototypes untouched. But you can extend them as below.
// you have to explicitly extend String.prototype
Base64.extendString();
// once extended, you can do the following
'dankogai'.toBase64(); // ZGFua29nYWk=
'小飼弾'.toBase64(); // 5bCP6aO85by+
'小飼弾'.toBase64(true); // 5bCP6aO85by-
'小飼弾'.toBase64URI(); // 5bCP6aO85by- an alias of .toBase64(true)
'小飼弾'.toBase64URL(); // 5bCP6aO85by- an alias of .toBase64URI()
'ZGFua29nYWk='.fromBase64(); // dankogai
'5bCP6aO85by+'.fromBase64(); // 小飼弾
'5bCP6aO85by-'.fromBase64(); // 小飼弾
'5bCP6aO85by-'.toUint8Array();// UTF-8 bytes of '小飼弾'// you have to explicitly extend Uint8Array.prototype
Base64.extendUint8Array();
// once extended, you can do the following
u8s.toBase64(); // 'ZGFua29nYWk='
u8s.toBase64URI(); // 'ZGFua29nYWk'
u8s.toBase64URL(); // 'ZGFua29nYWk' an alias of .toBase64URI()// extend all at once
Base64.extendBuiltins().decode() vs .atob (and .encode() vs btoa())Suppose you have:
var pngBase64 =
"iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mNkYAAAAAYAAjCB0C8AAAAASUVORK5CYII=";Which is a Base64-encoded 1x1 transparent PNG, DO NOT USE Base64.decode(pngBase64). Use Base64.atob(pngBase64) instead. Base64.decode() decodes to UTF-8 string while Base64.atob() decodes to bytes, which is compatible to browser built-in atob(). The same rule applies to the opposite direction.
Or even better, Base64.toUint8Array(pngBase64).
base64.mjs and base64.js are compiled from base64.ts via Rollup.base64.js is ES5-compatible again (hence IE11-compatible).js-base64 switch to ES2015 module so it is no longer compatible with legacy browsers like IE (see above)