Need help with node-re2?
Click the “chat” button below for chat support from the developer who created it, or find similar developers for support.

About the developer

293 Stars 39 Forks Other 422 Commits 6 Opened issues


node.js bindings for RE2: fast, safe alternative to backtracking regular expression engines.

Services available


Need anything else?

Contributors list

node-re2 NPM version

This project provides bindings for RE2: fast, safe alternative to backtracking regular expression engines written by Russ Cox. To learn more about RE2, start with an overview Regular Expression Matching in the Wild. More resources can be found at his Implementing Regular Expressions page.

's regular expression language is almost a superset of what is provided by
(see Syntax), but it lacks two features: backreferences and lookahead assertions. See below for more details.

object emulates standard
making it a practical drop-in replacement in most cases.
is extended to provide
-based regular expression methods as well. To help to convert
objects to
its constructor can take
directly honoring all properties.

It can work with node.js buffers directly reducing overhead on recoding and copying characters, and making processing/parsing long files fast.

All documentation can be found in this README and in the wiki.

Why use node-re2?

The built-in Node.js regular expression engine can run in exponential time with a special combination: - A vulnerable regular expression - "Evil input"

This can lead to what is known as a Regular Expression Denial of Service (ReDoS). To tell if your regular expressions are vulnerable, you might try the one of these projects: - rxxr2 - safe-regex

However, neither project is perfect.

node-re2 can protect your Node.js application from ReDoS. node-re2 makes vulnerable regular expression patterns safe by evaluating them in

instead of the built-in Node.js regex engine.

Standard features

object can be created just like

Supported properties:

Supported methods:

Starting with 1.6.0 following well-known symbol-based methods are supported (see Symbols):

It allows to use

instances on strings directly, just like
var re = new RE2("1");
"213".match(re);        // [ '1', index: 1, input: '213' ]
"213".search(re);       // 1
"213".replace(re, "+"); // 2+3
"213".split(re);        // [ '2', '3' ]

Starting with 1.8.0 named groups are supported.


Shortcut construction

object can be created from a regular expression:
var re1 = new RE2(/ab*/ig); // from a RegExp object
var re2 = new RE2(re1);     // from another RE2 object



defines four more methods that can use regular expressions.
provides them as methods exchanging positions of a string, and a regular expression:

Starting with 1.6.0, these methods added as well-known symbol-based methods to be used transparently with ES6 string/regex machinery.


In order to support

directly, most methods can accept buffers instead of strings. It speeds up all operations. Following signatures are supported:
  • re2.exec(buf)
  • re2.test(buf)
  • re2.match(buf)
  • re2.split(buf[, limit])
  • re2.replace(buf, replacer)

Differences with their string-based versions:

  • All buffers are assumed to be encoded as UTF-8 (ASCII is a proper subset of UTF-8).
  • Instead of strings they return
    objects, even in composite objects. A buffer can be converted to a string with
  • All offsets and lengths are in bytes, rather than characters (each UTF-8 character can occupy from 1 to 4 bytes). This way users can properly slice buffers without costly recalculations from characters to bytes.


is used with a replacer function, the replacer can return a buffer, or a string. But all arguments (except for an input object) will be strings, and an offset will be in characters. If you prefer to deal with buffers and byte offsets in a replacer function, set a property
on the function:
function strReplacer(match, offset, input) {
    // typeof match == "string"
    return "<= " + offset + " characters|";

RE2("б").replace("абв", strReplacer); // "а<= 1 characters|в"

function bufReplacer(match, offset, input) { // typeof match == "string" return "<= " + offset + " bytes|"; } bufReplacer.useBuffers = true;

RE2("б").replace("абв", bufReplacer); // "а<= 2 bytes|в"

This feature works for string and buffer inputs. If a buffer was used as an input, its output will be returned as a buffer too, otherwise a string will be returned.

Calculate length

Two functions to calculate string sizes between UTF-8 and UTF-16 are exposed on

  • RE2.getUtf8Length(str)
    — calculates a buffer size in bytes to encode a UTF-16 string as a UTF-8 buffer.
  • RE2.getUtf16Length(buf)
    — calculates a string size in characters to encode a UTF-8 buffer as a UTF-16 string.

JavaScript supports UCS-2 strings with 16-bit characters, while node.js 0.11 supports full UTF-16 as a default string.


Starting 1.8.0 property

emulates the same property of
, meaning that it can be used to create an identical
instance. Sometimes, for troubleshooting purposes, a user wants to inspect a
translated source. It is available as a read-only property called

Unicode warning level

engine always works in the Unicode mode. In most cases either there is no difference or the Unicode mode is actually preferred. But sometimes a user wants a tight control over their regular expressions. For those cases, there is a static string property

Regular expressions in the Unicode mode work as usual. But if a regular expression lacks the Unicode flag, it is always added silently.

const x = /./;
x.flags; // ''
const y = new RE2(x);
y.flags; // 'u'

In the latter case

can do following actions depending on
  • 'nothing'
    (the default): no warnings or notifications of any kind, a regular expression will be created with
  • 'warnOnce'
    : warns exactly once the very first time, a regular expression will be created with
    • Assigning this value resets an internal flag, so
      will warn once again.
  • 'warn'
    : warns every time, a regular expression will be created with
  • 'throw'
    : throws a
    every time.
  • All other warning level values are silently ignored on asignment leaving the previous value unchanged.

Warnings and exceptions help to audit an application for stray non-Unicode regular expressions.

How to install


npm install --save re2

While the project is known to work with other package managers, it is not guaranteed nor tested. For example, yarn is known to fail in some scenarios (see this Wiki article).

Precompiled artifacts

When installing re2 the install script attempts to download a prebuilt artifact for your system from the Github releases. The download location can be overridden by setting the

environment variable as seen in the install script.

If all attempts to download the prebuilt artifact for your system fails the script attempts to built re2 locally on your machine using node-gyp.

How to use

It is used just like a

var RE2 = require("re2");

// with default flags var re = new RE2("a(b*)"); var result = re.exec("abbc"); console.log(result[0]); // "abb" console.log(result[1]); // "bb"

result = re.exec("aBbC"); console.log(result[0]); // "a" console.log(result[1]); // ""

// with explicit flags re = new RE2("a(b*)", "i"); result = re.exec("aBbC"); console.log(result[0]); // "aBb" console.log(result[1]); // "Bb"

// from regular expression object var regexp = new RegExp("a(b*)", "i"); re = new RE2(regexp); result = re.exec("aBbC"); console.log(result[0]); // "aBb" console.log(result[1]); // "Bb"

// from regular expression literal re = new RE2(/a(b*)/i); result = re.exec("aBbC"); console.log(result[0]); // "aBb" console.log(result[1]); // "Bb"

// from another RE2 object var rex = new RE2(re); result = rex.exec("aBbC"); console.log(result[0]); // "aBb" console.log(result[1]); // "Bb"

// shortcut result = new RE2("ab*").exec("abba");

// factory result = RE2("ab*").exec("abba");

Limitations (things RE2 does not support)

consciously avoids any regular expression features that require worst-case exponential time to evaluate. These features are essentially those that describe a Context-Free Language (CFL) rather than a Regular Expression, and are extensions to the traditional regular expression language because some people don't know when enough is enough.

The most noteworthy missing features are backreferences and lookahead assertions. If your application uses these features, you should continue to use

. But since these features are fundamentally vulnerable to ReDoS, you should strongly consider replacing them.

will throw a
if you try to declare a regular expression using these features. If you are evaluating an externally-provided regular expression, wrap your
declarations in a try-catch block. It allows to use
, when
misses a feature:
var re = /(a)+(b)*/;
try {
  re = new RE2(re);
  // use RE2 as a drop-in replacement
} catch (e) {
  // suppress an error, and use
  // the original RegExp
var result = re.exec(sample);

In addition to these missing features,

also behaves somewhat differently from the built-in regular expression engine in corner cases.


doesn't support backreferences, which are numbered references to previously matched groups, like so:
, and so on. Example of backrefrences:
/(cat|dog)\1/.test("catcat"); // true
/(cat|dog)\1/.test("dogdog"); // true
/(cat|dog)\1/.test("catdog"); // false
/(cat|dog)\1/.test("dogcat"); // false

Lookahead assertions

doesn't support lookahead assertions, which are ways to allow a matching dependent on subsequent contents.
/abc(?=def)/; // match abc only if it is followed by def
/abc(?!def)/; // match abc only if it is not followed by def

Mismatched behavior

and the built-in regex engines disagree a bit. Before you switch to
, verify that your regular expressions continue to work as expected. They should do so in the vast majority of cases.

Here is an example of a case where they may not:

var RE2  = require("../re2");

var pattern = '(?:(a)|(b)|(c))+';

var built_in = new RegExp(pattern); var re2 = new RE2(pattern);

var input = 'abc';

var bi_res = built_in.exec(input); var re2_res = re2.exec(input);

console.log('bi_res: ' + bi_res); // prints: bi_res: abc,,,c console.log('re2_res : ' + re2_res); // prints: re2_res : abc,a,b,c


always works in the Unicode mode. See
above for more details on how to control warnings about this feature.

Release history

  • 1.16.0 Updated the compiler (thx, Sergei Dyshel), updated GYP.
  • 1.15.9 Updated deps.
  • 1.15.8 Updated deps.
  • 1.15.7 Updated deps.
  • 1.15.6 Technical release: less dependencies for the build.
  • 1.15.5 Updated deps. Fixed a
    -related problem (thx malte-v).
  • 1.15.4 Updated deps. Fixed a yarn-related bug (thx Michael Kriese).
  • 1.15.3 Extracted caching artifacts to separate packages. Added support for
    environment variable for precompiled artifact download during installation.
  • 1.15.2 Added
    target for precompiled images (thx Uzlopak).
  • 1.15.1 Refreshed dependencies, updated the verification check on installation, general maintenance.
  • 1.15.0 Fix for multiline expressions (thx Frederic Rudman),
    now, updated deps.
  • 1.14.0 New delivery mechanism for binary artifacts (thx Brandon Kobel for the idea and the research) + minor fix to eliminate warnings on Windows.
  • 1.13.1 Fix for Windows builds.
  • 1.13.0 Got rid of a single static variable to support multithreading.
  • 1.12.1 Updated re2 to the latest version.
  • 1.12.0 Updated the way
    objects are constructed.
  • 1.11.0 Updated the way to initialize the extension (thx BannerBomb).
  • 1.10.5 Bugfix for optional groups (thx Josh Yudaken), the latest version of re2.
  • 1.10.4 Technical release: even better TypeScript types (thx Louis Brann).
  • 1.10.3 Technical release: missing reference to TS types (thx Jamie Magee).
  • 1.10.2 Technical release: added TypeScript types (thx Jamie Magee).
  • 1.10.1 Updated re2 to the latest version (thx Jamie Magee), dropped Node 6.
  • 1.10.0 Added back support for Node 6 and Node 8. Now Node 6-12 is supported.
  • 1.9.0 Refreshed dependencies to support Node 12. Only versions 10-12 are supported now (
    restrictions). For older versions use
    [email protected]
  • 1.8.4 Refreshed dependencies, removed
    to compile on Windows.
  • 1.8.3 Refreshed dependencies, removed suppression of some warnings.
  • 1.8.2 Bugfix to support the null prototype for groups. Thx Exter-N.
  • 1.8.1 Bugfix for better source escaping.
  • 1.8.0 Clarified Unicode support, added
    flag, added named groups — thx Exter-N! Bugfixes — thx Barak Amar.
  • 1.7.0 Implemented
    + bug fixes + more tests. Thx Exter-N.
  • 1.6.2 Bugfix for a prototype access. Thx Exter-N.
  • 1.6.1 Returned support for node 4 LTS. Thx Kannan Goundan.
  • 1.6.0 Added well-known symbol-based methods of ES6. Refreshed NAN.
  • 1.5.0 Bugfixes, error checks, better docs. Thx Jamie Davis, and omg.
  • 1.4.1 Minor corrections in README.
  • 1.4.0 Use re2 as a git submodule. Thx Ben James.
  • 1.3.3 Refreshed dependencies.
  • 1.3.2 Updated references in README (re2 was moved to github).
  • 1.3.1 Refreshed dependencies, new Travis-CI config.
  • 1.3.0 Upgraded NAN to 1.6.3, now we support node.js 0.10.36, 0.12.0, and io.js 1.3.0. Thx Reid Burke!
  • 1.2.0 Documented getUtfXLength() functions. Added support for
  • 1.1.1 Minor corrections in README.
  • 1.1.0 Buffer-based API is public. Unicode is fully supported.
  • 1.0.0 Implemented all
    methods, and all relevant
  • 0.9.0 The initial public release.



We use cookies. If you continue to browse the site, you agree to the use of cookies. For more information on our use of cookies please see our Privacy Policy.