Element (formerly known as Vector and Riot) is a Matrix web client built using the Matrix React SDK.
Element has several tiers of support for different environments:
- Supported
- Definition: Issues actively triaged, regressions block the release
- Last 2 major versions of Chrome, Firefox, Safari, and Edge on desktop OSes
- Latest release of official Element Desktop app on desktop OSes
- Desktop OSes means macOS, Windows, and Linux versions for desktop devices that are actively supported by the OS vendor and receive security updates
- Experimental
- Definition: Issues accepted, regressions do not block the release
- Element as an installed PWA via current stable version of Chrome, Firefox, and Safari
- Mobile web for current stable version of Chrome, Firefox, and Safari on Android, iOS, and iPadOS
- Not supported
- Definition: Issues only affecting unsupported environments are closed
- Everything else
For accessing Element on an Android or iOS device, we currently recommend the native apps element-android and element-ios.
The easiest way to test Element is to just use the hosted copy at https://app.element.io.
The develop
branch is continuously deployed to https://develop.element.io
for those who like living dangerously.
To host your own copy of Element, the quickest bet is to use a pre-built released version of Element:
- Download the latest version from https://github.com/vector-im/element-web/releases
- Untar the tarball on your web server
- Move (or symlink) the
element-x.x.x
directory to an appropriate name - Configure the correct caching headers in your webserver (see below)
- If desired, copy
config.sample.json
toconfig.json
and edit it as desired. See the configuration docs for details. - Enter the URL into your browser and log into Element!
Releases are signed using gpg and the OpenPGP standard, and can be checked against the public key located at https://packages.riot.im/element-release-key.asc.
Note that for the security of your chats will need to serve Element over HTTPS. Major browsers also do not allow you to use VoIP/video chats over HTTP, as WebRTC is only usable over HTTPS. There are some exceptions like when using localhost, which is considered a secure context and thus allowed.
To install Element as a desktop application, see Running as a desktop app below.
We do not recommend running Element from the same domain name as your Matrix homeserver. The reason is the risk of XSS (cross-site-scripting) vulnerabilities that could occur if someone caused Element to load and render malicious user generated content from a Matrix API which then had trusted access to Element (or other apps) due to sharing the same domain.
We have put some coarse mitigations into place to try to protect against this situation, but it's still not good practice to do it in the first place. See element-hq#1977 for more details.
Unless you have special requirements, you will want to add the following to your web server configuration when hosting Element Web:
- The
X-Frame-Options: SAMEORIGIN
header, to prevent Element Web from being framed and protect from clickjacking. - The
frame-ancestors 'none'
directive to yourContent-Security-Policy
header, as the modern replacement forX-Frame-Options
(though both should be included since not all browsers support it yet, see this). - The
X-Content-Type-Options: nosniff
header, to disable MIME sniffing. - The
X-XSS-Protection: 1; mode=block;
header, for basic XSS protection in legacy browsers.
If you are using nginx, this would look something like the following:
add_header X-Frame-Options SAMEORIGIN;
add_header X-Content-Type-Options nosniff;
add_header X-XSS-Protection "1; mode=block";
add_header Content-Security-Policy "frame-ancestors 'none'";
Note: In case you are already setting a Content-Security-Policy
header
elsewhere, you should modify it to include the frame-ancestors
directive
instead of adding that last line.
Element is a modular webapp built with modern ES6 and uses a Node.js build system. Ensure you have the latest LTS version of Node.js installed.
Using yarn
instead of npm
is recommended. Please see the Yarn install
guide if you do not have it already.
- Install or update
node.js
so that yournode
is at least v10.x. - Install
yarn
if not present already. - Clone the repo:
git clone https://github.com/vector-im/element-web.git
. - Switch to the element-web directory:
cd element-web
. - Install the prerequisites:
yarn install
.- If you're using the
develop
branch, then it is recommended to set up a proper development environment (see Setting up a dev environment below). Alternatively, you can use https://develop.element.io - the continuous integration release of the develop branch.
- If you're using the
- Configure the app by copying
config.sample.json
toconfig.json
and modifying it. See the configuration docs for details. yarn dist
to build a tarball to deploy. Untaring this file will give a version-specific directory containing all the files that need to go on your web server.
Note that yarn dist
is not supported on Windows, so Windows users can run yarn build
,
which will build all the necessary files into the webapp
directory. The version of Element
will not appear in Settings without using the dist script. You can then mount the
webapp
directory on your web server to actually serve up the app, which is
entirely static content.
Element can also be run as a desktop app, wrapped in Electron. You can download a pre-built version from https://element.io/get-started or, if you prefer, build it yourself.
To build it yourself, follow the instructions at https://github.com/vector-im/element-desktop.
Many thanks to @aviraldg for the initial work on the Electron integration.
Other options for running as a desktop app:
- @asdf:matrix.org points out that you can use nativefier and it just works(tm)
yarn global add nativefier
nativefier https://app.element.io/
The configuration docs show how to override the desktop app's default settings if desired.
The Docker image can be used to serve element-web as a web server. The easiest way to use it is to use the prebuilt image:
docker run -p 80:80 vectorim/element-web
To supply your own custom config.json
, map a volume to /app/config.json
. For example,
if your custom config was located at /etc/element-web/config.json
then your Docker command
would be:
docker run -p 80:80 -v /etc/element-web/config.json:/app/config.json vectorim/element-web
To build the image yourself:
git clone https://github.com/vector-im/element-web.git element-web
cd element-web
git checkout master
docker build .
If you're building a custom branch, or want to use the develop branch, check out the appropriate element-web branch and then run:
docker build -t \
--build-arg USE_CUSTOM_SDKS=true \
--build-arg REACT_SDK_REPO="https://github.com/matrix-org/matrix-react-sdk.git" \
--build-arg REACT_SDK_BRANCH="develop" \
--build-arg JS_SDK_REPO="https://github.com/matrix-org/matrix-js-sdk.git" \
--build-arg JS_SDK_BRANCH="develop" \
.
The provided element-web docker image can also be run from within a Kubernetes cluster. See the Kubernetes example for more details.
Element supports a variety of settings to configure default servers, behaviour, themes, etc. See the configuration docs for more details.
Some features of Element may be enabled by flags in the Labs
section of the settings.
Some of these features are described in labs.md.
Element requires the following URLs not to be cached, when/if you are serving Element from your own webserver:
/config.*.json
/i18n
/home
/sites
/index.html
Before attempting to develop on Element you must read the developer guide
for matrix-react-sdk
, which
also defines the design, architecture and style for Element too.
Before starting work on a feature, it's best to ensure your plan aligns well with our vision for Element. Please chat with the team in #element-dev:matrix.org before you start so we can ensure it's something we'd be willing to merge.
You should also familiarise yourself with the "Here be Dragons" guide to the tame & not-so-tame dragons (gotchas) which exist in the codebase.
The idea of Element is to be a relatively lightweight "skin" of customisations on
top of the underlying matrix-react-sdk
. matrix-react-sdk
provides both the
higher and lower level React components useful for building Matrix communication
apps using React.
After creating a new component you must run yarn reskindex
to regenerate
the component-index.js
for the app (used in future for skinning).
Please note that Element is intended to run correctly without access to the public internet. So please don't depend on resources (JS libs, CSS, images, fonts) hosted by external CDNs or servers but instead please package all dependencies into Element itself.
Much of the functionality in Element is actually in the matrix-react-sdk
and
matrix-js-sdk
modules. It is possible to set these up in a way that makes it
easy to track the develop
branches in git and to make local changes without
having to manually rebuild each time.
First clone and build matrix-js-sdk
:
git clone https://github.com/matrix-org/matrix-js-sdk.git
pushd matrix-js-sdk
yarn link
yarn install
popd
Then similarly with matrix-react-sdk
:
git clone https://github.com/matrix-org/matrix-react-sdk.git
pushd matrix-react-sdk
yarn link
yarn link matrix-js-sdk
yarn install
popd
Finally, build and start Element itself:
git clone https://github.com/vector-im/element-web.git
cd element-web
yarn link matrix-js-sdk
yarn link matrix-react-sdk
yarn install
yarn start
Wait a few seconds for the initial build to finish; you should see something like:
[element-js] <s> [webpack.Progress] 100%
[element-js]
[element-js] ℹ 「wdm」: 1840 modules
[element-js] ℹ 「wdm」: Compiled successfully.
Remember, the command will not terminate since it runs the web server and rebuilds source files when they change. This development server also disables caching, so do NOT use it in production.
Configure the app by copying config.sample.json
to config.json
and
modifying it. See the configuration docs for details.
Open http://127.0.0.1:8080/ in your browser to see your newly built Element.
Note: The build script uses inotify by default on Linux to monitor directories
for changes. If the inotify limits are too low your build will fail silently or with
Error: EMFILE: too many open files
. To avoid these issues, we recommend a watch limit
of at least 128M
and instance limit around 512
.
You may be interested in issues #15750 and #15774 for further details.
To set a new inotify watch and instance limit, execute:
sudo sysctl fs.inotify.max_user_watches=131072
sudo sysctl fs.inotify.max_user_instances=512
sudo sysctl -p
If you wish, you can make the new limits permanent, by executing:
echo fs.inotify.max_user_watches=131072 | sudo tee -a /etc/sysctl.conf
echo fs.inotify.max_user_instances=512 | sudo tee -a /etc/sysctl.conf
sudo sysctl -p
When you make changes to matrix-react-sdk
or matrix-js-sdk
they should be
automatically picked up by webpack and built.
If you add or remove any components from the Element skin, you will need to rebuild
the skin's index by running, yarn reskindex
.
If any of these steps error with, file table overflow
, you are probably on a mac
which has a very low limit on max open files. Run ulimit -Sn 1024
and try again.
You'll need to do this in each new terminal you open before building Element.
There are a number of application-level tests in the tests
directory; these
are designed to run in a browser instance under the control of
karma. To run them:
- Make sure you have Chrome installed (a recent version, like 59)
- Make sure you have
matrix-js-sdk
andmatrix-react-sdk
installed and built, as above yarn test
The above will run the tests under Chrome in a headless
mode.
You can also tell karma to run the tests in a loop (every time the source
changes), in an instance of Chrome on your desktop, with yarn test-multi
. This also gives you the option of running the tests in 'debug'
mode, which is useful for stepping through the tests in the developer tools.
See matrix-react-sdk how to run the end-to-end tests.
To add a new translation, head to the translating doc.
For a developer guide, see the translating dev doc.
We strive to completely cover all applicable issues with these core labels:
-
Type — Every issue is assigned a type:
- T-Defect: Bugs, crashes, hangs, vulnerabilities, or other reported problems
- T-Enhancement: New features, changes in functionality, performance boosts, user-facing improvements
- T-Task: Refactoring, enabling or disabling functionality, other engineering tasks
- T-Other: Questions, user support, anything else
-
Severity — All issues labeled
T-Defect
are also assigned a severity:- S-Critical: Prevents work, causes data loss, affects many users, and/or has no workaround
- S-Major: Severely degrades major functionality or product features, with no satisfactory workaround
- S-Minor: Impairs non-critical functionality, or suitable workarounds exist
- S-Tolerable: Purely cosmetic or low / no impact to users
-
Priority — All issues which are not
T-Other
are assigned a priority: -
Area — Most issues are assigned one or several "areas" using one of the many
A-
prefixed labels, e.g.A-Composer
orA-Spaces
. Each area label maps to a group of features or portion of the UI surface in the app.
We have a handful of other labels which are added on an as-needed basis, and not expected to be exhaustive:
-
Exceptions — Special flags for issues and pull requests:
- X-Needs-Info: This issue is blocked pending further information from the reporter
- X-Regression: Denotes things breaking which previously worked
- X-Release-Blocker: Issues which must be resolved before making a release
-
Easy / Help Wanted — Well-defined issues which are suitable for folks new to the codebase
-
A11y / Meta / I18n / Privacy / Security — Issues which fall under these conceptual themes (which apply to many software projects and are not specific to Element)
-
Sponsored — Used internally by Element to denote issues with external funding
We have reserved the Z-
prefix for ad hoc labels.
Any member of the core team is welcome to create labels beginning with Z-
for
any purpose, such as tracking personal areas of interest or providing a common
way to label cross-repo initiatives. The prefix avoids interference with the
project's main labels.