diff --git a/migrations/2024-03-17-104156_reset_schema/up.sql b/migrations/2024-03-17-104156_reset_schema/up.sql
index 4ee8218..805e578 100644
--- a/migrations/2024-03-17-104156_reset_schema/up.sql
+++ b/migrations/2024-03-17-104156_reset_schema/up.sql
@@ -97,7 +97,9 @@ CREATE TABLE recording_ensembles (
 CREATE TABLE tracks_new (
     track_id TEXT NOT NULL PRIMARY KEY,
     recording_id TEXT NOT NULL REFERENCES recordings(recording_id),
-    sequence_number INTEGER NOT NULL,
+    recording_index INTEGER NOT NULL,
+    medium_id TEXT REFERENCES mediums(medium_id),
+    medium_index INTEGER,
     path TEXT NOT NULL,
     created_at TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP,
     edited_at TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP,
@@ -112,6 +114,38 @@ CREATE TABLE track_works (
     PRIMARY KEY (track_id, work_id)
 );
 
+CREATE TABLE mediums_new (
+    medium_id TEXT NOT NULL PRIMARY KEY,
+    discid TEXT NOT NULL,
+    created_at TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP,
+    edited_at TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP,
+    last_used_at TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP,
+    last_played_at TIMESTAMP
+);
+
+CREATE TABLE albums (
+    album_id TEXT NOT NULL PRIMARY KEY,
+    name TEXT NOT NULL,
+    created_at TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP,
+    edited_at TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP,
+    last_used_at TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP,
+    last_played_at TIMESTAMP
+);
+
+CREATE TABLE album_recordings (
+    album_id TEXT NOT NULL REFERENCES albums(album_id),
+    recording_id TEXT NOT NULL REFERENCES recordings(recording_id),
+    sequence_number INTEGER NOT NULL,
+    PRIMARY KEY (album_id, recording_id)
+);
+
+CREATE TABLE album_mediums (
+    album_id TEXT NOT NULL REFERENCES albums(album_id),
+    medium_id TEXT NOT NULL REFERENCES mediums(medium_id),
+    sequence_number INTEGER NOT NULL,
+    PRIMARY KEY (album_id, medium_id)
+);
+
 INSERT INTO persons_new (person_id, name)
 SELECT id, json_set('{}', '$.generic', first_name || ' ' || last_name)
 FROM persons;
@@ -164,8 +198,8 @@ SELECT recording, ensemble, '28ff0aeb11c041a6916d93e9b4884eef', 0
 FROM performances
 WHERE ensemble IS NOT NULL;
 
-INSERT INTO tracks_new (track_id, recording_id, sequence_number, path)
-SELECT id, recording, "index", path
+INSERT INTO tracks_new (track_id, recording_id, recording_index, medium_id, medium_index, path)
+SELECT id, recording, "index", medium, source_index, path
 FROM tracks;
 
 INSERT INTO track_works (track_id, work_id, sequence_number)
@@ -175,6 +209,27 @@ FROM tracks
     JOIN work_parts ON recordings.work = work_parts.work
         AND tracks.work_parts = work_parts.part_index;
 
+INSERT INTO mediums_new (medium_id, discid)
+SELECT id, discid
+FROM mediums;
+
+INSERT INTO albums (album_id, name)
+SELECT id, json_set('{}', '$.generic', name)
+FROM mediums;
+
+-- This works because mediums, that become albums in this migration, keep their ID.
+INSERT INTO album_recordings (album_id, recording_id, sequence_number)
+SELECT medium, recording, ROW_NUMBER() OVER (PARTITION BY medium ORDER BY min_source_index)
+FROM (
+    SELECT medium, recording, MIN(source_index) AS min_source_index
+    FROM tracks GROUP BY medium, recording
+);
+
+-- See above
+INSERT INTO album_mediums (album_id, medium_id, sequence_number)
+SELECT id, id, 0
+FROM mediums;
+
 DROP TABLE persons;
 DROP TABLE instruments;
 DROP TABLE works;
@@ -191,4 +246,5 @@ ALTER TABLE instruments_new RENAME TO instruments;
 ALTER TABLE works_new RENAME TO works;
 ALTER TABLE recordings_new RENAME TO recordings;
 ALTER TABLE tracks_new RENAME TO tracks;
-ALTER TABLE ensembles_new RENAME TO ensembles;
\ No newline at end of file
+ALTER TABLE ensembles_new RENAME TO ensembles;
+ALTER TABLE mediums_new RENAME TO mediums;
\ No newline at end of file
diff --git a/src/db/models.rs b/src/db/models.rs
index d641d7d..3bbe304 100644
--- a/src/db/models.rs
+++ b/src/db/models.rs
@@ -260,7 +260,7 @@ impl Recording {
             .collect::<Result<Vec<Ensemble>>>()?;
 
         let tracks: Vec<Track> = tracks::table
-            .order(tracks::sequence_number)
+            .order(tracks::recording_index)
             .filter(tracks::recording_id.eq(&data.recording_id))
             .select(tables::Track::as_select())
             .load::<tables::Track>(connection)?
diff --git a/src/db/schema.rs b/src/db/schema.rs
index 6fdf29e..a847d60 100644
--- a/src/db/schema.rs
+++ b/src/db/schema.rs
@@ -1,5 +1,32 @@
 // @generated automatically by Diesel CLI.
 
+diesel::table! {
+    album_mediums (album_id, medium_id) {
+        album_id -> Text,
+        medium_id -> Text,
+        sequence_number -> Integer,
+    }
+}
+
+diesel::table! {
+    album_recordings (album_id, recording_id) {
+        album_id -> Text,
+        recording_id -> Text,
+        sequence_number -> Integer,
+    }
+}
+
+diesel::table! {
+    albums (album_id) {
+        album_id -> Text,
+        name -> Text,
+        created_at -> Timestamp,
+        edited_at -> Timestamp,
+        last_used_at -> Timestamp,
+        last_played_at -> Nullable<Timestamp>,
+    }
+}
+
 diesel::table! {
     ensemble_persons (ensemble_id, person_id, instrument_id) {
         ensemble_id -> Text,
@@ -31,6 +58,17 @@ diesel::table! {
     }
 }
 
+diesel::table! {
+    mediums (medium_id) {
+        medium_id -> Text,
+        discid -> Text,
+        created_at -> Timestamp,
+        edited_at -> Timestamp,
+        last_used_at -> Timestamp,
+        last_played_at -> Nullable<Timestamp>,
+    }
+}
+
 diesel::table! {
     persons (person_id) {
         person_id -> Text,
@@ -95,7 +133,9 @@ diesel::table! {
     tracks (track_id) {
         track_id -> Text,
         recording_id -> Text,
-        sequence_number -> Integer,
+        recording_index -> Integer,
+        medium_id -> Nullable<Text>,
+        medium_index -> Nullable<Integer>,
         path -> Text,
         created_at -> Timestamp,
         edited_at -> Timestamp,
@@ -143,6 +183,10 @@ diesel::table! {
     }
 }
 
+diesel::joinable!(album_mediums -> albums (album_id));
+diesel::joinable!(album_mediums -> mediums (medium_id));
+diesel::joinable!(album_recordings -> albums (album_id));
+diesel::joinable!(album_recordings -> recordings (recording_id));
 diesel::joinable!(ensemble_persons -> ensembles (ensemble_id));
 diesel::joinable!(ensemble_persons -> instruments (instrument_id));
 diesel::joinable!(ensemble_persons -> persons (person_id));
@@ -156,6 +200,7 @@ diesel::joinable!(recording_persons -> roles (role_id));
 diesel::joinable!(recordings -> works (work_id));
 diesel::joinable!(track_works -> tracks (track_id));
 diesel::joinable!(track_works -> works (work_id));
+diesel::joinable!(tracks -> mediums (medium_id));
 diesel::joinable!(tracks -> recordings (recording_id));
 diesel::joinable!(work_instruments -> instruments (instrument_id));
 diesel::joinable!(work_instruments -> works (work_id));
@@ -164,9 +209,13 @@ diesel::joinable!(work_persons -> roles (role_id));
 diesel::joinable!(work_persons -> works (work_id));
 
 diesel::allow_tables_to_appear_in_same_query!(
+    album_mediums,
+    album_recordings,
+    albums,
     ensemble_persons,
     ensembles,
     instruments,
+    mediums,
     persons,
     recording_ensembles,
     recording_persons,
diff --git a/src/db/tables.rs b/src/db/tables.rs
index 98c4d50..5421685 100644
--- a/src/db/tables.rs
+++ b/src/db/tables.rs
@@ -129,7 +129,9 @@ pub struct RecordingEnsemble {
 pub struct Track {
     pub track_id: String,
     pub recording_id: String,
-    pub sequence_number: i32,
+    pub recording_index: i32,
+    pub medium_id: Option<String>,
+    pub medium_index: Option<i32>,
     pub path: String,
     pub created_at: NaiveDateTime,
     pub edited_at: NaiveDateTime,
@@ -144,3 +146,41 @@ pub struct TrackWork {
     pub work_id: String,
     pub sequence_number: i32,
 }
+
+#[derive(Insertable, Queryable, Selectable, Clone, Debug)]
+#[diesel(check_for_backend(Sqlite))]
+pub struct Medium {
+    pub medium_id: String,
+    pub discid: String,
+    pub created_at: NaiveDateTime,
+    pub edited_at: NaiveDateTime,
+    pub last_used_at: NaiveDateTime,
+    pub last_played_at: Option<NaiveDateTime>,
+}
+
+#[derive(Insertable, Queryable, Selectable, Clone, Debug)]
+#[diesel(check_for_backend(Sqlite))]
+pub struct Album {
+    pub album_id: String,
+    pub name: TranslatedString,
+    pub created_at: NaiveDateTime,
+    pub edited_at: NaiveDateTime,
+    pub last_used_at: NaiveDateTime,
+    pub last_played_at: Option<NaiveDateTime>,
+}
+
+#[derive(Insertable, Queryable, Selectable, Clone, Debug)]
+#[diesel(check_for_backend(Sqlite))]
+pub struct AlbumRecording {
+    pub album_id: String,
+    pub recording_id: String,
+    pub sequence_number: i32,
+}
+
+#[derive(Insertable, Queryable, Selectable, Clone, Debug)]
+#[diesel(check_for_backend(Sqlite))]
+pub struct AlbumMedium {
+    pub album_id: String,
+    pub medium_id: String,
+    pub sequence_number: i32,
+}
\ No newline at end of file