This is basically Fluent Emoji from Microsoft.

The folder and file structure has been processed to easily request the URL of a specific emoji using their unicode value.

Additionally, every emoji has been rendered in WEBP.

The formats available:

• WEBP: 32x32, 64x64, 128x128
• SVG: source, and minified.

Featuring both Fluent Emoji families: 3D (Color) and Flat.

Everything has been processed in the assets folder. To replicate the results or update it:

1. Fetch the icons from Microsoft's Fluent Emoji repo (~130mb).

git clone https://github.com/microsoft/fluentui-emoji.git

2. Make sure you have:

• having Python 3 (latest version)
• Imagemagick in the system
• inkscape. Imagemagick uses it as the backend for SVG processing. Otherwise, it will fallback to a less capable SVG processor, which fails at rendering the SVGs.
• and node with NPM to install svgo.

Install svgo:

npm -g install svgo

or use Yarn:

yarn global add svgo

3. Then:

python convert.py
./svg-processing.sh

convert.py will transform the folder and file structure from Microsoft's Fluent Emoji repo to the one we need.

svg-processing.sh minifies the SVGs and renders WEBP alternatives. It may take a long, long time. Like, 1 hour or so. It doesn't use all the cores because of performance loss due to overhead. You could find the optimal number of jobs for your CPU via the parallel -j3 line, where 3 is the number of jobs. Compare the differents ETA times shown by the script when running at different number of jobs. This probably can be automated but I don't want to :).