╔══──--∙·.      .·∙-──────────────────────══════╗
║ ▪   ▄▄·        ▐  ▄ ▌  ▐·▄▄▄▄▄▄ ▪  ▐  ▄·▄· ▄▌ ║
│ ██ ▐█ ▌▪▪     •█▌▐█▪█·▐▌  •██  ██ •█▌▐█ █▪█▌  ║
│ ▐█·██ ▄▄ ▄█▀▄ ▐█▐▐▌ ▐ █ •▄•▐█.▪▐█·▐█▐▐▌ ▐█▪   │
│ ▐█▌▐█ █▌▐█▌.▐▌█▌▐█▌ ▐█▌    ▐█▌·▐█ █▌▐█▌ ▐█·.  │
│ ▀▀▀·▀▀▀  ▀█▄▀▪▀  █▪. ▀     ▀▀▀ ▀▀ ▀  █▪  ▀ •  │
└──────--∙·    ·∙-───────────────by─ViP-DeLeTe──┘

Pure JS character encoding conversion as an ECMAScript Module (ESM). Auto-Generated from https://encoding.spec.whatwg.org and https://www.unicode.org/Public/MAPPINGS.

About

iconv-tiny.mjs is written from scratch and designed to address all issues and implement all features in the iconv-lite backlog.
It is an ES-module and should work in all modern browsers and NodeJS that supports ESM, TextDecoder, and TextEncoder.

Features

1. Zero dependencies.
2. Browser support out-of-the-box.
3. Tree shaking and other ESM-related benefits.
4. Table mappings and default characters overrides.
5. Native TextDecoder for some encodings.
6. encodeInto(Uint8Array) and byteLength(string) functions
7. Typescript friendly.
8. Tiny: 1 file with dbcs tables, ~30KB in gzip.

Installation

npm install iconv-tiny

or use CDN:

<script type="importmap">
  {
    "imports": {
      "iconv-tiny": "https://unpkg.com/iconv-tiny@1.4.0/dist/iconv-tiny.mjs"
    }
  }
</script>

Basic API

API is very close to iconv-lite API, see iconv-tiny.d.mts.

import { canonicalize, createIconv, encodings, aliases } from "iconv-tiny";

// Create iconv
const iconv = createIconv(encodings, aliases);

// Convert from an Uint8Array to a js string.
str = iconv.decode(new Uint8Array([0x68, 0x65, 0x6c, 0x6c, 0x6f]), "win1251");

// Convert from a js string to an Uint8Array.
buf = iconv.encode("Sample input string", "win1251"); // Uint8Array

// Check if encoding is supported
aliases.split(/,| /).map(canonicalize).includes(canonicalize("us-ascii"));

or import encodings individually to enable ESM bundlers to remove unused encodings and reduce the final bundle size.

import { CP1251 } from "iconv-tiny";

const cp1251 = CP1251.create();
const buf = cp1251.encode("Век живи — век учись.");
const str = cp1251.decode(buf);

Streaming API

import { UTF16LE } from "iconv-tiny";

// Create encoding
const utf16 = UTF16LE.create();

// Create decoder, it works like TextDecoder with {stream: true} option.
const decoder = utf16.getDecoder();

// Decode a fragment
const part = decoder.write(new Uint8Array([0x3d, 0xd8, 0x0a]));

// Decode the next fragment
const str = decoder.write(new Uint8Array([0xde])); // 😊

// Finish stream decoding
const tail = decoder.end();

See more examples.

Supported encodings (singlebyte & doublebyte encodings, Unicode):

1. ISO-8859: 1,2,3,4,5,6,7,8,9,10,11,13,14,15,16
2. EBCDIC: Cp037, Cp424, Cp500, Cp875, Cp1026
3. DOS: Cp437, Cp737, Cp775, Cp850, Cp852, Cp855, Cp857, Cp860, Cp861, Cp862, Cp863, Cp864, Cp865, Cp866, Cp869, Cp874
4. WINDOWS: Cp1250, Cp1251, Cp1252, Cp1253, Cp1254, Cp1255, Cp1256, Cp1257, Cp1258
5. MAC: CYRILLIC, GREEK, ICELAND, LATIN2, ROMAN, TURKISH
6. MISC: ATARIST, CP856, Cp1006, KOI8-R, KOI8-U, KZ1048, NEXTSTEP
7. OTHER: US-ASCII
8. UNICODE: UTF-8, UTF-16, UTF-32
9. Japanese: JIS-0201, SHIFT-JIS, CP932

All encodings are generated automatically from https://www.unicode.org/Public/MAPPINGS with a few additional mappings for CP875 (0xE1 -> 0x20AF, 0xEC -> 0x037A, 0xFC -> 0x20AC) and CP1255 (0xCA -> 0x05BA), see mappings

iconv-tiny output is identical to iconv-lite output, see tests/regression.test.mjs.

Encoding/decoding speed in NodeJS

Comparison with iconv-lite module (Core i7-7500U CPU @ 2.7GHz, Node v24.2.0). Note: your results may vary, so please always check on your hardware.

operation          iconv-lite@0.7.2   iconv-tiny@1.4.0
------------------------------------------------------
encode('win1251')     ~270 Mb/s          ~270 Mb/s
decode('win1251')     ~120 Mb/s          ~220 Mb/s

iconv-lite is NodeJS oriented and use specific API like Buffer and native NodeJS encodings.
iconv-tiny is Browser oriented and use Uint8Array and native TextDecoder, and TextEncoder. Also it supports encodeInto(Uint8Array) and byteLength(string).

BOM Handling

1. UTF-8, UTF-16LE, UTF-16BE, UTF-32LE, UTF-32BE:
   • Decoding: BOM is stripped by default, unless overridden by stripBOM: false option.
   • Encoding: No BOM added, unless overridden by addBOM: true option.
2. UTF-16 is an alias of UTF-16LE
3. UTF-32 is an alias of UTF-32LE

Browser Testing

https://vip-delete.github.io/iconv-tiny/tests/whatwg.html

NodeJS Testing & Coverage

$ git clone https://github.com/vip-delete/iconv-tiny.git
$ cd iconv-tiny
$ npm test

$ # To view performance:
$ node tests\perf-test-cp1251.mjs
$ node tests\perf-test-unicode.mjs

$ # To view test coverage:
$ npm run coverage
-------------|---------|----------|---------|---------|-------------------
File         | % Stmts | % Branch | % Funcs | % Lines | Uncovered Line #s
-------------|---------|----------|---------|---------|-------------------
All files    |   97.86 |    95.55 |   96.66 |   97.93 |
 commons.mjs |   93.83 |       95 |   97.29 |    94.2 | 36-41,53-56
 dbcs.mjs    |   97.02 |    89.47 |     100 |   96.87 | 27,134,143
 iconv.mjs   |     100 |      100 |     100 |     100 |
 mapped.mjs  |     100 |      100 |     100 |     100 |
 native.mjs  |     100 |      100 |     100 |     100 |
 sbcs.mjs    |     100 |     90.9 |     100 |     100 | 36-66
 types.mjs   |     100 |      100 |      50 |     100 |
 unicode.mjs |     100 |      100 |     100 |     100 |
 utf16.mjs   |     100 |      100 |     100 |     100 |
 utf32.mjs   |   98.76 |     97.5 |     100 |   98.76 | 224
 utf8.mjs    |     100 |      100 |     100 |     100 |
-------------|---------|----------|---------|---------|-------------------

Commands

Build:

npm run build

UI Demo:

npm run dev

Compile (optional, you need JDK 21 or higher installed):

npm run compile

Demo

https://vip-delete.github.io/iconv-tiny/