mediaproxy

Proxy Images and other files through itself. Also can generate Discord Embeds out of given websites, via OpenGraph or <meta> tags or other embedders.

This is an optional component of Litecord.

This is designed to run on a separate machine, so the server running the litecord instance does not have to deal with other websites.

Discord does not use this software.

Installation

Requirements:

• Python 3.7+
• pipenv

mediaproxy does not require a database. The filesystem is used as cache.

Download the code

$ git clone https://gitlab.com/litecord/mediaproxy.git && cd mediaproxy

Install packages

$ pipenv install

Configuration

$ cp config.example.ini config.ini
# edit config.ini as wanted

Running

It is preferred to run mediaproxy on the port 5002, while Litecord and its websocket run on ports 5000 and 5001 respectively.

$ pipenv run hypercorn run:app --bind 0.0.0.0:5002

It is recommended to run mediaproxy behind NGINX. You can use the nginx.conf file at the root of the repository as a template.

Usage

external websites use the following syntax in paths:

/<scheme>/<path>

path is composed of <host>/<url_path>. scheme can only be http or https.

• the /img/ scope proxies an external HTTP website's multimedia resource.
• the /meta scope gives metadata about an external image, such as width and height, used for embed generation in litecord
• the /embed/ scope gives a discord embed representing the url. depending of the url's format it can be routed to a different embedder, such as the OGP (OpenGraph Protocol) embedder, <meta> embedder, or XKCD embedder, etc.

Errors

• Requesting images or metadata from URLs that aren't images will return a 405 Method Not Allowed response.
• Embeds that fail to be generated will return null.

Example

if you want to get a certain image's raw data as a response body, http://blah.com/img/http/website.com/photo.png

replace img with meta for image metadata, and img with embed to generate an embed.