diff --git a/README.md b/README.md
index 07fe936..4f67bb6 100644
--- a/README.md
+++ b/README.md
@@ -1,3 +1,55 @@
-# backend
+# Jos
 
-The backend, written using Python and Flask.
\ No newline at end of file
+Jos is one of my oldest project ideas, and the one that's definitely stuck
+around the longest.
+
+The idea is pretty straightforward: I like music, and I'd like to download any
+music I want from YouTube using
+[YoutubeDL](https://github.com/ytdl-org/youtube-dl)
+([mirror](https://git.roosens.me/mirrors/youtube-dl)). I've written dozens of
+scripts that automate some parts of this process, such as looking up the
+correct video, or searching for metadata using the
+[Spotify API](https://developer.spotify.com/documentation/web-api/). The goal
+of this project is to be the final "script", housing everything I need to grow
+my music collection.
+
+## Design of the project
+
+I've decided to create Jos using the classic frontend-backend combo: a backend
+(written in Python using Flask) serves a RESTful API that does all the hard
+work. We then put a lovely frontend in front of it (Vue 3) that interacts with
+this API to make a functional website.
+
+Ever since I had my Web Development course (taught by
+[Ruben Verborgh](https://ruben.verborgh.org/)), I've been very interested in
+developing high-performing and practical web applications using RESTful API's.
+I know developing two separate applications requires more work, but it allows
+for more flexibility, e.g. it allows me to swap out the backend for a
+completely new implementation,
+[in a different language](https://git.roosens.me/Chewing_Bever/fej).
+
+## Plans
+
+### API
+
+* Search engine for high-quality music videos (using the `"auto-generated"`
+  trick.)
+* Wrapper layer around the Spotify API for searching for tracks/albums/artists
+* Return a JSON dump containing all metadata & urls you need to download a
+  track/album
+    This allows the API to be used by non-authenticated users as well (e.g. we
+    could write a Python client that uses the API, but downloads locally.)
+* Download a track/album (should use some sort of authentification probably)
+
+### Frontend
+
+* Search for tracks/albums/artists
+* Edit metadata & select best video
+* Tell server to download track/album
+    * Live updates per track on how far the download is (or just wether or not
+      it's done)
+*  file explorer for music collection 
+*  metadata editor for collection
+* Analysis of existing collection
+    * Find incomplete albums & suggest missing tracks to download
+    * Detect invalid metadata
diff --git a/ideation.md b/ideation.md
deleted file mode 100644
index 16cd4fe..0000000
--- a/ideation.md
+++ /dev/null
@@ -1,26 +0,0 @@
-# Ideation
-
-This file just serves as a central point for all the ideas I have for this
-project.
-
-## Used software/services
-
-* Spotify API to look up album information
-* YouTubeDL to download the videos
-* Something to set the mp3 tags, not sure about this one yet
-
-## API Design
-
-### Endpoints
-
-* `GET /search?q=`
-    * search for albums & tracks, given a search query
-    * returns a list of albums/tracks (basically the same as the API, just
-      cleaned up a bit)
-* `GET /albums?ids=`: get detailed information about an album/a list of albums.
-  I'm not actually sure if I need this one.
-* `GET /tracks?ids=`: get detailed information about a track/a list of tracks.
-* `GET /tracks/videos?id=`: given a track, search YouTube for videos about the
-  track. Returns a list of youtube ids
-* `POST /tracks/download`: download a track. The data should contain the wanted
-  metadata in the file & the id of the youtube video