A modern messaging platform for Message Push、IM、IoT etc, based on MQTT protocol
Switch branches/tags
Nothing to show
Clone or download
Type Name Latest commit message Commit time
Failed to load latest commit information.
broker update Sep 6, 2018
demos updaate vendor Jun 30, 2018
proto fix some minor issues Aug 16, 2018
sdks delete meq-website,used mafanr website instead Sep 9, 2018
vendor update vendor Jun 30, 2018
.gitignore resolve #76, each user in chat topic will have its own unread count now Jun 14, 2018
LICENSE refactor route of cluster, refactor go client sdk Jun 14, 2018
Readme.md update Sep 6, 2018
ReadmeCn.md update Sep 6, 2018
logo.png update logo Jun 29, 2018


MeQ [mi:kju]

A modern messaging platform for Message Push、IM、Group Chatting、IoT etc, based on MQTT protocol. MeQ is written in pure go and standard library,nearly no messy dependencies. so you can easily deploy a standalone binary in linux、unix、macos、windows.

Our goal is to be the best messaging platform in the world.

Project status

Still in active developing, 2018-08-16

Example of Chatting room

Install the foundationdb client(required!)

Mac OS: https://www.foundationdb.org/downloads/5.1.7/macOS/installers/FoundationDB-5.1.7.pkg

CentOS7: https://www.foundationdb.org/downloads/5.1.7/rhel7/installers/foundationdb-clients-5.1.7-1.el7.x86_64.rpm

CentOS6: https://www.foundationdb.org/downloads/5.1.7/rhel6/installers/foundationdb-clients-5.1.7-1.el6.x86_64.rpm

Other Os: https://www.foundationdb.org/download/ (Please choose version 5.1.7)

Install foundationdb server(optional,only needed when you use fdb store mode)

CentOS7: https://www.foundationdb.org/downloads/5.1.7/rhel7/installers/foundationdb-server-5.1.7-1.el7.x86_64.rpm

Start MeQ

> go get github.com/mafanr/meq
> cd $GOPATH/src/github.com/mafanr/meq/broker
> go run main.go

Start example

> cd ../demos/chatting
> npm install
> npm run dev

Chatting now!

Open your browser and accsess http://localhost:8080 in two seperate pages,in one page input username A,in other page input username B, then you can start chatting!

More(persistent storage)

Broker uses memory storage by default, if you want to use persistent storage, please install foundationDB server V5.1.7 , and then edit the broker.yaml,change store.engine from memory to fdb.

Now Status

The version is Alpha V0.5.0. The website,documents and a group chat demo will soon be possible to access(before 2018.6.30)


Extremely performant

  • Written in go, as fast as c/c++ network programing
  • Zero allocation and copy
  • Hot path and algorithm optimized

Easy and efficient

  • Easy to study and use
  • Efficient when you need to develop on meq,because the nice design and code
  • Detailed docs and examples


  • Robust is one of our most important goal throughout the developing
  • Default message persistent,also you can implement the persistent inteface in your way.

Various scenario supported

  • Message push
  • Group chatting
  • IM
  • IoT messaging
  • Real time Web interaction,like dashboard、online collaboration etc

Group chat

  • Join and leave the group
  • Each user has a separate unread message number,e.g. you have 97 unread messages in ethereum-welcom group
  • You can retrive your message even after sended
  • History messages playback
  • Query all the users or online users in the group
  • Every message will store only once, all the users share the message

Advanced feature

  • Topic wild match, e.g. you can push to /china/+/city1, then all the china province which has city named city1, will receive the message.
  • Mqtt and websocket,tcp etc
  • Low latency, message push deliver to user as soon as possible
  • Fertile administrator features and ui
  • Monitor and message trace

Why MeQ?

There were already many messaging platform out there, but none of them can satisfy all the messaging scenario, e.g., if you want build a online chatting platform, you need to build it from scratch, nearly none of the existing platform can satisfy your need. So �here meq comes, cover your messaging scenario in the most easy way.


Want to be part of the meq project? Great! All are welcome! We will get there quicker together :) Whether you find a bug, have a great feature request or you fancy owning a task from the road map, you can leave a issue.

Contributors(Sort by contribution)

Performance(early stage)

In this benchmark, I use the memory engine, all is done in my macbook pro laptop.

  • A client with 5 goroutine can publish 2700K messages to meq per second
  • A client with 5 goroutine can consume 2000K messages from meq per second