Robust and scalable Jabber / XMPP Instant Messaging platform
Erlang C Shell Graphviz (DOT) Perl Python CSS
Pull request Compare This branch is 13 commits ahead, 4794 commits behind esl:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.


esl-ejabberd is Erlang Solutions' robust and efficient XMPP server aimed at large installations. It utilizes Erlang/OTP's hallmark features - fault tolerance and suitability for server software working in highly demanding conditions.

Its home at GitHub is

Quickstart guide

For a quick start just download the pre-built package that suits your platform.

Main differences from the parent project

This project began its life as a fork of ejabberd v.2.1.8 and later underwent some major cleanup, refactorization and optimization.

Major steps performed:

  • bringing the project source tree to compliance with OTP project structure recommendations,
  • swapping autotools for the Erlang community-standard build tool rebar,
  • removal of obsolete and/or rarely used modules to reduce maintenance burden,
  • reduction of runtime memory consumption by refactoring the code to use Erlang's binary data type for string manipulation and storage instead of operating on linked lists of characters,
  • functional test coverage of the system according to corresponding RFCs and XEPs.

How to build

  1. Requirements.

    To compile ejabberd you need:

    • GNU Make,
    • GCC,
    • Libexpat 1.95 or higher,
    • Erlang/OTP R15B or higher,
    • Reltool 0.5.4 or higher,
    • OpenSSL 0.9.8 or higher, for STARTTLS, SASL and SSL encryption,
    • Zlib 1.2.3 or higher for Stream Compression support (XEP-0138). Optional,
    • PAM library. Optional. For Pluggable Authentication Modules (PAM).
  2. Compiling on UNIX-like systems.

    To compile ejabberd, go to the main repo directory $REPO and execute the command ($ stands for the shell prompt):

    $ make


    $ ./rebar compile

    To generate ejabberd release:

    $ make rel


    $ ./rebar generate

    These commands will generate a self-contained OTP system image in the project's rel/ejabberd subdirectory. The contents of that directory are as follows:

    • rel/ejabberd/bin - startup/administration scripts,
    • rel/ejabberd/etc - configuration files,
    • rel/ejabberd/lib - ejabberd binary, header and runtime files,
    • rel/ejabberd/var - spool directory,
    • rel/ejabberd/log - log file directory,
    • rel/ejabberd/releases - release files directory.
  3. Running esl-ejabberd.

    To run ejabberd from the project tree after compiling it, change to $REPO/rel/ejabberd.

    There you can use the ejabberd command line administration script to start and stop ejabberd. For example:

    $ bin/ejabberd start

    will start the server.

    You can also run the server in interactive mode:

    $ bin/ejabberd live

    There's also a tool called ejabberdctl allowing you to perform some operations on a running instance, e.g.:

    $ bin/ejabberdctl status
    The node ejabberd@localhost is started with status: started
    ejabberd 2.1.8 is running in that node
  4. Building the testing target and running tests.

    For testing purposes there's a different make target available:

    $ make testrel

    which will generate releases in $REPO/dev/ and prepare them for testing and generating coverage reports.

    To run the tests (from project's root directory, i.e. $REPO):

    $ cd test
    $ make cover_test

    The test results will show up in the console and a coverage report will be generated in $REPO/test/ct_report/cover.html.

esl-ejabberd documentation notice

esl-ejabberd, being a fork of ejabberd, is currently in a phase of rapid development. Because of that the documentation found in the doc/ subdirectory of the source tree, while mostly relevant, may be inaccurate.

The main reason is that some rarely useful features were removed from the repository or are still waiting to be brought up to the Erlang Solutions standards.

Want to get in touch with us?

In case of any suggestions, questions or any random thoughts on this project, please feel free to contact us by the standard GitHub ways or at