Mastodon

Mastodon is a free, open-source social network server. A decentralized alternative to commercial platforms, it avoids the risks of a single company monopolizing your communication. Anyone can run Mastodon and participate in the social network seamlessly.

An alternative implementation of the GNU social project. Based on ActivityStreams, Webfinger, PubsubHubbub and Salmon.

Click on the screenshot to watch a demo of the UI:

Focus of the project on a clean REST API and a good user interface. Ruby on Rails is used for the back-end, while React.js and Redux are used for the dynamic front-end. A static front-end for public resources (profiles and statuses) is also provided.

If you would like, you can support the development of this project on Patreon. Alternatively, you can donate to this BTC address: 17j2g7vpgHhLuXhN4bueZFCvdxxieyRVWd

Resources

• List of Mastodon instances
• Use this tool to find Twitter friends on Mastodon
• API overview
• How to use the API via cURL/oAuth
• Frequently Asked Questions
• List of apps

Features

• Fully interoperable with GNU social and any OStatus platform Whatever implements Atom feeds, ActivityStreams, Salmon, PubSubHubbub and Webfinger is part of the network
• Real-time timeline updates See the updates of people you're following appear in real-time in the UI via WebSockets
• Federated thread resolving If someone you follow replies to a user unknown to the server, the server fetches the full thread so you can view it without leaving the UI
• Media attachments like images and WebM Upload and view images and WebM videos attached to the updates
• OAuth2 and a straightforward REST API Mastodon acts as an OAuth2 provider so 3rd party apps can use the API, which is RESTful and simple
• Background processing for long-running tasks Mastodon tries to be as fast and responsive as possible, so all long-running tasks that can be delegated to background processing, are
• Deployable via Docker You don't need to mess with dependencies and configuration if you want to try Mastodon, if you have Docker and Docker Compose the deployment is extremely easy

Configuration

• LOCAL_DOMAIN should be the domain/hostname of your instance. This is absolutely required as it is used for generating unique IDs for everything federation-related
• LOCAL_HTTPS set it to true if HTTPS works on your website. This is used to generate canonical URLs, which is also important when generating and parsing federation-related IDs

Consult the example configuration file, .env.production.sample for the full list. Among other things you need to set details for the SMTP server you are going to use.

Requirements

• Ruby
• Node.js
• PostgreSQL
• Redis
• Nginx

Running with Docker and Docker-Compose

The project now includes a Dockerfile and a docker-compose.yml. You need to turn .env.production.sample into .env.production with all the variables set before you can:

docker-compose build

And finally

docker-compose up -d

As usual, the first thing you would need to do would be to run migrations:

docker-compose run --rm web rails db:migrate

And since the instance running in the container will be running in production mode, you need to pre-compile assets:

docker-compose run --rm web rails assets:precompile

The container has two volumes, for the assets and for user uploads. The default docker-compose.yml maps them to the repository's public/assets and public/system directories, you may wish to put them somewhere else. Likewise, the PostgreSQL and Redis images have data containers that you may wish to map somewhere where you know how to find them and back them up.

Note: The --rm option for docker-compose will remove the container that is created to run a one-off command after it completes. As data is stored in volumes it is not affected by that container clean-up.

Tasks

• rake mastodon:media:clear removes uploads that have not been attached to any status after a while, you would want to run this from a periodic cronjob
• rake mastodon:push:clear unsubscribes from PuSH notifications for remote users that have no local followers. You may not want to actually do that, to keep a fuller footprint of the fediverse or in case your users will soon re-follow
• rake mastodon:push:refresh re-subscribes PuSH for expiring remote users, this should be run periodically from a cronjob and quite often as the expiration time depends on the particular hub of the remote user
• rake mastodon:feeds:clear_all removes all timelines, which forces them to be re-built on the fly next time a user tries to fetch their home/mentions timeline. Only for troubleshooting
• rake mastodon:feeds:clear removes timelines of users who haven't signed in lately, which allows to save RAM and improve message distribution. This is required to be run periodically so that when they login again the regeneration process will trigger

Running any of these tasks via docker-compose would look like this:

docker-compose run --rm web rake mastodon:media:clear

Updating

This approach makes updating to the latest version a real breeze.

git pull

To pull down the updates, re-run

docker-compose build

And finally,

docker-compose up -d

Which will re-create the updated containers, leaving databases and data as is. Depending on what files have been updated, you might need to re-run migrations and asset compilation.

Deployment without Docker

Docker is great for quickly trying out software, but it has its drawbacks too. If you prefer to run Mastodon without using Docker, refer to the production guide for examples, configuration and instructions.

Deployment on Heroku (experimental)

Mastodon can theoretically run indefinitely on a free Heroku app. It should be noted this has limited testing and could have unpredictable results.

1. Click the above button.
2. Fill in the options requested.

• You can use a .herokuapp.com domain, which will be simple to set up, or you can use a custom domain. If you want a custom domain and HTTPS, you will need to upgrade to a paid plan (to use Heroku's SSL features), or set up CloudFlare who offer free "Flexible SSL" (note: CloudFlare have some undefined limits on WebSockets. So far, no one has reported hitting concurrent connection limits).
• You will want Amazon S3 for file storage. The only exception is for development purposes, where you may not care if files are not saaved. Follow a guide online for creating a free Amazon S3 bucket and Access Key, then enter the details.
• If you want your Mastodon to be able to send emails, configure SMTP settings here (or later). Consider using Mailgun or similar, who offer free plans that should suit your interests.

3. Deploy! The app should be set up, with a working web interface and database. You can change settings and manage versions from the Heroku dashboard.

Development with Vagrant

A quick way to get a development environment up and running is with Vagrant. You will need recent versions of Vagrant and VirtualBox installed.

Install the latest version for your operating systems, and then run:

vagrant plugin install vagrant-hostsupdater

This is optional, but will update your 'hosts' file when you start the virtual machine, allowing you to access the site at http://mastodon.dev (instead of http://localhost:3000).

To create and provision a new virtual machine for Mastodon development:

git clone git@github.com:tootsuite/mastodon.git
cd mastodon
vagrant up

Running vagrant up for the first time will run provisioning, which will:

• Download the Ubuntu 14.04 base image, if there isn't already a copy on your machine
• Create a new VirtualBox virtual machine from that image
• Run the provisioning script (located inside the Vagrantfile), which installs the system packages, Ruby gems, and JS modules required for Mastodon

Once this has completed, the virtual machine will start a rails process. You can then access your development site at http://mastodon.dev (or at http://localhost:3000 if you haven't installed vagrants-hostupdater). Any changes you make should be reflected on the server instantly. To set environment variables, copy .env.production.sample to .env.vagrant and make changes as required.

When you are finished with your session, run vagrant halt to stop the VM. Next time, running vagrant up should boot the VM, and skip provisioning.

If you no longer need your environment, or if things have gone terribly wrong, running vagrant destroy will delete the virtual machine (after which, running vagrant up will create a new one, and run provisioning).

Contributing

You can open issues for bugs you've found or features you think are missing. You can also submit pull requests to this repository. This section may be updated with more details in the future.

IRC channel: #mastodon on irc.freenode.net

Extra credits

• The Emoji One pack has been used for the emojis
• The error page image courtesy of Dopatwo