diff --git a/README.md b/README.md
new file mode 100644
index 0000000..1b0ea0c
--- /dev/null
+++ b/README.md
@@ -0,0 +1,16 @@
+# mscz2video
+
+Generate a video from a `.mscz` (MuseScore) file
+
+The goal of this repo is to **fully automate** this. It however still requires a bit of manual work.
+It is targeted towards a choir (SATB) but can be used for other instruments as well.
+
+## Steps to generate a video
+
+1. Get - or create - the mscz file. You will need two versions of it : one for the visual score, which will have to look clean, you can combine voices etc. The other for generating the audios ; all voices need to be on a separate staff, but you don't need lyrics for example.
+2. Generate images `python src/generate_images.py <visual_score.mscz> <output/image>`
+3. Create an ffconcat file with the timestamps between two images, the playing capability of MuseScore (with its timer) will help you accomplish this (see `src/images_to_video.md`)
+4. Concatenate the images to create a video `ffmpeg -y -f concat -i concat.txt -vf "fps=5,format=yuv420p" -vsync vfr output.webm`
+5. Generate audios `python src/mscz_to_mp3.py <audio_score.mscz> out`
+6. If necessary, combine instruments with `src/superpose_audios.sh`
+7. Combine audio and video using `src/set_audio.sh`