diff --git a/tests/metadata/test_album_type_normalization.py b/tests/metadata/test_album_type_normalization.py
index c628ef53..ac806c1e 100644
--- a/tests/metadata/test_album_type_normalization.py
+++ b/tests/metadata/test_album_type_normalization.py
@@ -130,3 +130,19 @@ class TestLegacyBuilderAlbumTypeAltKeys:
         a stray value from poisoning the path template."""
         info = self._build({'type': 'Mixtape'})
         assert info['album_type'] == 'album'
+
+    def test_type_track_collision_defaults_to_album(self):
+        """`type` is a generic key name — many dict shapes use it
+        for entity discrimination (Deezer track responses carry
+        `type='track'`, Spotify search results use it for
+        `artist`/`track`/`album`/etc). If a caller hands us a dict
+        that has `type='track'` (track data passed by mistake, or a
+        merged shape where `type` discriminates entity rather than
+        album category), the normalizer must treat it as unknown and
+        default to `album` — not silently classify the result as some
+        bogus 'track' folder."""
+        info = self._build({'type': 'track'})
+        assert info['album_type'] == 'album'
+
+        info = self._build({'type': 'artist'})
+        assert info['album_type'] == 'album'
diff --git a/webui/static/helper.js b/webui/static/helper.js
index 75075e09..83274182 100644
--- a/webui/static/helper.js
+++ b/webui/static/helper.js
@@ -3416,7 +3416,7 @@ const WHATS_NEW = {
     '2.5.2': [
         // --- post-release patch work on the 2.5.2 line — entries hidden by _getLatestWhatsNewVersion until the build version bumps ---
         { date: 'Unreleased — 2.5.2 patch work' },
-        { title: '$albumtype Folder Template Now Splits EPs / Singles For Non-Spotify Sources', desc: 'discord report (cal): downloading an artist\'s discography with `$albumtype` in the path template put every release under `Album/` regardless of actual type — eps, singles, all dumped into the album folder. trace: the legacy duck-typed album-info builder at `core/metadata/album_tracks.py:_build_album_info_legacy` only checked the `album_type` key. spotify uses `album_type` (lowercase) so spotify discographies worked. but deezer\'s api uses `record_type`, tidal uses `type` (uppercase ALBUM/EP/SINGLE), and some flattened musicbrainz shapes use `primary-type` — none of those matched, all defaulted to `album`. fix: widen the legacy lookup to check `album_type` / `record_type` / `type` / `primary-type` and route the value through a new pure `_normalize_album_type` helper that lowercases + validates against the canonical token set (`album` / `single` / `ep` / `compilation`) and falls back to `album` for unknowns. typed-converter path for spotify / deezer / itunes / discogs / musicbrainz / hydrabase / qobuz unchanged — they were already correct. tidal users were the main offender (no typed converter for dict-shaped tidal data). 24 new tests pin: case-insensitive normalization for each canonical type, compilation preserved (spotify supports it), unknown values default to album, defensive against none / empty / non-string inputs, multi-key precedence (`album_type` wins over `record_type`), each known source shape produces correct token.', page: 'tools' },
+        { title: '$albumtype Folder Template Now Splits EPs / Singles For Non-Spotify Sources', desc: 'discord report (cal): downloading an artist\'s discography with `$albumtype` in the path template put every release under `Album/` regardless of actual type — eps, singles, all dumped into the album folder. trace: the legacy duck-typed album-info builder at `core/metadata/album_tracks.py:_build_album_info_legacy` only checked the `album_type` key. spotify uses `album_type` (lowercase) so spotify discographies worked. but deezer\'s api uses `record_type`, tidal uses `type` (uppercase ALBUM/EP/SINGLE), and some flattened musicbrainz shapes use `primary-type` — none of those matched, all defaulted to `album`. fix: widen the legacy lookup to check `album_type` / `record_type` / `type` / `primary-type` and route the value through a new pure `_normalize_album_type` helper that lowercases + validates against the canonical token set (`album` / `single` / `ep` / `compilation`) and falls back to `album` for unknowns. typed-converter path for spotify / deezer / itunes / discogs / musicbrainz / hydrabase / qobuz unchanged — they were already correct. tidal users were the main offender (no typed converter for dict-shaped tidal data). 25 new tests pin: case-insensitive normalization for each canonical type, compilation preserved (spotify supports it), unknown values default to album, defensive against none / empty / non-string inputs, multi-key precedence (`album_type` wins over `record_type`), each known source shape produces correct token, generic `type='track'` / `type='artist'` collision case defaults to album rather than poisoning the path.', page: 'tools' },
     ],
     '2.5.1': [
         // --- May 12, 2026 — 2.5.1 release ---