Skip to content
This repository


Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP

node.js iconv bindings - text recoding for fun and profit!


Text recoding in JavaScript for fun and profit!

node-iconv may or may not work on Windows. Please try it and report any issues you have.

Installing with npm

$ npm install iconv

Note that you do not need to have a copy of libiconv installed to use this module.

Compiling from source

$ git clone git://
$ node-gyp configure build
$ npm install .


Encode from one character encoding to another:

// convert from UTF-8 to ISO-8859-1
var Buffer = require('buffer').Buffer;
var Iconv  = require('iconv').Iconv;
var assert = require('assert');

var iconv = new Iconv('UTF-8', 'ISO-8859-1');
var buffer = iconv.convert('Hello, world!');
var buffer2 = iconv.convert(new Buffer('Hello, world!'));
assert.equals(buffer.inspect(), buffer2.inspect());
// do something useful with the buffers

A simple ISO-8859-1 to UTF-8 conversion TCP service:

var net = require('net');
var Iconv = require('iconv').Iconv;
var server = net.createServer(function(conn) {
  var iconv = new Iconv('latin1', 'utf-8');
console.log('Listening on tcp://');

Look at test/test-basic.js and test/test-stream.js for more examples and node-iconv's behaviour under error conditions.


Things to keep in mind when you work with node-iconv.

Chunked data

Say you are reading data in chunks from a HTTP stream. The logical input is a single document (the full POST request data) but the physical input will be spread over several buffers (the request chunks).

You must accumulate the small buffers into a single large buffer before performing the conversion. If you don't, you will get unexpected results with multi-byte and stateful character sets like UTF-8 and ISO-2022-JP.

The above only applies when you are calling Iconv#convert() yourself. If you use the streaming interface, node-iconv takes care of stitching partial character sequences together again.

Dealing with untranslatable characters

Characters are not always translatable to another encoding. The UTF-8 string "ça va が", for example, cannot be represented in plain 7-bits ASCII without some loss of fidelity.

By default, node-iconv throws EILSEQ when untranslatabe characters are encountered but this can be customized. Quoting the iconv_open(3) man page:

When  the  string  "//TRANSLIT"  is appended to tocode, transliteration is
activated. This means that when a character cannot be represented in the
target character set, it can be approximated through one or several
similarly looking characters.

When the string "//IGNORE" is appended to tocode, characters that cannot be
represented in the target character set will be silently discarded.

Example usage:

var iconv = new Iconv('UTF-8', 'ASCII');
iconv.convert('ça va'); // throws EILSEQ

var iconv = new Iconv('UTF-8', 'ASCII//IGNORE');
iconv.convert('ça va'); // returns "a va"

var iconv = new Iconv('UTF-8', 'ASCII//TRANSLIT');
iconv.convert('ça va'); // "ca va"

var iconv = new Iconv('UTF-8', 'ASCII//TRANSLIT//IGNORE');
iconv.convert('ça va が'); // "ca va "


EINVAL is raised when the input ends in a partial character sequence. This is a feature, not a bug.

Something went wrong with that request. Please try again.