diff --git a/web_server.py b/web_server.py
index 0e55ea72..8be5cc66 100644
--- a/web_server.py
+++ b/web_server.py
@@ -2764,6 +2764,63 @@ def add_cache_headers(response, cache_duration=300):
     response.headers['Pragma'] = 'cache'
     return response
 
+
+_enhanced_search_cache = collections.OrderedDict()
+_enhanced_search_cache_lock = threading.Lock()
+_ENHANCED_SEARCH_CACHE_TTL = 600
+_ENHANCED_SEARCH_CACHE_MAX_ENTRIES = 100
+
+
+def _get_enhanced_search_cache_key(query):
+    """Build a cache key that follows the current metadata/search configuration."""
+    normalized_query = (query or '').strip().lower()
+
+    try:
+        active_server = config_manager.get_active_media_server()
+    except Exception:
+        active_server = 'unknown'
+
+    try:
+        fallback_source = _get_metadata_fallback_source()
+    except Exception:
+        fallback_source = 'unknown'
+
+    try:
+        hydrabase_active = _is_hydrabase_active()
+    except Exception:
+        hydrabase_active = False
+
+    return (normalized_query, active_server, fallback_source, hydrabase_active)
+
+
+def _get_cached_enhanced_search_response(cache_key):
+    """Return a cached enhanced-search response if it is still fresh."""
+    now = time.time()
+    with _enhanced_search_cache_lock:
+        entry = _enhanced_search_cache.get(cache_key)
+        if not entry:
+            return None
+
+        if now - entry['timestamp'] < _ENHANCED_SEARCH_CACHE_TTL:
+            _enhanced_search_cache.move_to_end(cache_key)
+            return entry['data']
+
+        _enhanced_search_cache.pop(cache_key, None)
+        return None
+
+
+def _set_cached_enhanced_search_response(cache_key, response_data):
+    """Store an enhanced-search response in the short-lived in-memory cache."""
+    with _enhanced_search_cache_lock:
+        _enhanced_search_cache[cache_key] = {
+            'timestamp': time.time(),
+            'data': response_data,
+        }
+        _enhanced_search_cache.move_to_end(cache_key)
+
+        while len(_enhanced_search_cache) > _ENHANCED_SEARCH_CACHE_MAX_ENTRIES:
+            _enhanced_search_cache.popitem(last=False)
+
 # --- Background Download Monitoring (GUI Parity) ---
 class WebUIDownloadMonitor:
     """
@@ -8514,6 +8571,7 @@ def enhanced_search():
     """
     data = request.get_json()
     query = data.get('query', '').strip()
+    cache_key = _get_enhanced_search_cache_key(query)
 
     empty_source = {"artists": [], "albums": [], "tracks": [], "available": False}
 
@@ -8528,6 +8586,11 @@ def enhanced_search():
             "metadata_source": "spotify"
         })
 
+    cached_response = _get_cached_enhanced_search_response(cache_key)
+    if cached_response is not None:
+        logger.info(f"Enhanced search cache hit for: '{query}'")
+        return jsonify(cached_response)
+
     logger.info(f"Enhanced search initiated for: '{query}'")
 
     try:
@@ -8547,6 +8610,23 @@ def enhanced_search():
                 "image_url": image_url
             })
 
+        # Very short queries are usually broad enough that remote metadata searches
+        # just add latency without improving the result quality much. Keep them local.
+        if len(query) < 3:
+            fb_source = _get_metadata_fallback_source()
+            response_data = {
+                "db_artists": db_artists,
+                "spotify_artists": [],
+                "spotify_albums": [],
+                "spotify_tracks": [],
+                "metadata_source": fb_source,
+                "primary_source": fb_source,
+                "alternate_sources": [],
+                "sources": {},
+            }
+            _set_cached_enhanced_search_response(cache_key, response_data)
+            return jsonify(response_data)
+
         # ── Determine primary source and search it synchronously ──
         primary_source = "spotify"
         primary_results = empty_source
@@ -8576,7 +8656,7 @@ def enhanced_search():
             # Search using the user's configured primary metadata source
             fb_source = _get_metadata_fallback_source()
             try:
-                primary_results = _enhanced_search_source(query, _get_metadata_fallback_client())
+                primary_results = _enhanced_search_source(query, _get_metadata_fallback_client(), fb_source)
                 primary_source = fb_source
             except Exception as e:
                 logger.debug(f"Primary source ({fb_source}) search failed: {e}")
@@ -8585,7 +8665,7 @@ def enhanced_search():
             if primary_results is empty_source and fb_source != 'spotify':
                 if spotify_client and spotify_client.is_spotify_authenticated():
                     try:
-                        primary_results = _enhanced_search_source(query, spotify_client)
+                        primary_results = _enhanced_search_source(query, spotify_client, "spotify")
                         primary_source = "spotify"
                     except Exception as e:
                         logger.debug(f"Spotify fallback search failed: {e}")
@@ -8615,7 +8695,7 @@ def enhanced_search():
                      f"{len(primary_results['tracks'])} tracks | "
                      f"Alt sources available: {alternate_sources}")
 
-        return jsonify({
+        response_data = {
             # Backward compat — same shape as before
             "db_artists": db_artists,
             "spotify_artists": primary_results["artists"],
@@ -8625,66 +8705,95 @@ def enhanced_search():
             # New multi-source data
             "primary_source": primary_source,
             "alternate_sources": alternate_sources,
-        })
+        }
+        _set_cached_enhanced_search_response(cache_key, response_data)
+        return jsonify(response_data)
 
     except Exception as e:
         logger.error(f"Enhanced search error: {e}")
         return jsonify({"error": str(e)}), 500
 
 
-def _enhanced_search_source(query, client):
+def _search_metadata_source_kind(client, query, kind, source_name=None):
+    """Search one result type from a metadata source and normalize it."""
+    source_label = source_name or type(client).__name__
+
+    if kind == "artists":
+        artists = []
+        try:
+            artist_objs = client.search_artists(query, limit=10)
+            for artist in artist_objs:
+                artists.append({
+                    "id": artist.id,
+                    "name": artist.name,
+                    "image_url": artist.image_url,
+                    "external_urls": artist.external_urls or {},
+                })
+        except Exception as e:
+            logger.debug(f"Artist search failed for {source_label}: {e}")
+        return artists
+
+    if kind == "albums":
+        albums = []
+        try:
+            album_objs = client.search_albums(query, limit=10)
+            for album in album_objs:
+                artist_name = ', '.join(album.artists) if album.artists else 'Unknown Artist'
+                albums.append({
+                    "id": album.id,
+                    "name": album.name,
+                    "artist": artist_name,
+                    "image_url": album.image_url,
+                    "release_date": album.release_date,
+                    "total_tracks": album.total_tracks,
+                    "album_type": album.album_type,
+                    "external_urls": album.external_urls or {},
+                })
+        except Exception as e:
+            logger.warning(f"Album search failed for {source_label}: {e}", exc_info=True)
+        return albums
+
+    if kind == "tracks":
+        tracks = []
+        try:
+            track_objs = client.search_tracks(query, limit=10)
+            for track in track_objs:
+                artist_name = ', '.join(track.artists) if track.artists else 'Unknown Artist'
+                tracks.append({
+                    "id": track.id,
+                    "name": track.name,
+                    "artist": artist_name,
+                    "album": track.album,
+                    "duration_ms": track.duration_ms,
+                    "image_url": track.image_url,
+                    "release_date": track.release_date,
+                    "external_urls": track.external_urls or {},
+                })
+        except Exception as e:
+            logger.warning(f"Track search failed for {source_label}: {e}", exc_info=True)
+        return tracks
+
+    raise ValueError(f"Unknown metadata search kind: {kind}")
+
+
+def _enhanced_search_source(query, client, source_name=None):
     """Search a single metadata source and return normalized results dict."""
-    artists = []
-    albums = []
-    tracks = []
+    results = {"artists": [], "albums": [], "tracks": []}
+    with ThreadPoolExecutor(max_workers=3) as executor:
+        futures = {
+            executor.submit(_search_metadata_source_kind, client, query, "artists", source_name): "artists",
+            executor.submit(_search_metadata_source_kind, client, query, "albums", source_name): "albums",
+            executor.submit(_search_metadata_source_kind, client, query, "tracks", source_name): "tracks",
+        }
+        for future in as_completed(futures):
+            kind = futures[future]
+            try:
+                results[kind] = future.result()
+            except Exception as e:
+                logger.warning(f"{kind.title()} search failed for {source_name or type(client).__name__}: {e}", exc_info=True)
+                results[kind] = []
 
-    try:
-        artist_objs = client.search_artists(query, limit=10)
-        for artist in artist_objs:
-            artists.append({
-                "id": artist.id,
-                "name": artist.name,
-                "image_url": artist.image_url,
-                "external_urls": artist.external_urls or {},
-            })
-    except Exception as e:
-        logger.debug(f"Artist search failed for {type(client).__name__}: {e}")
-
-    try:
-        album_objs = client.search_albums(query, limit=10)
-        for album in album_objs:
-            artist_name = ', '.join(album.artists) if album.artists else 'Unknown Artist'
-            albums.append({
-                "id": album.id,
-                "name": album.name,
-                "artist": artist_name,
-                "image_url": album.image_url,
-                "release_date": album.release_date,
-                "total_tracks": album.total_tracks,
-                "album_type": album.album_type,
-                "external_urls": album.external_urls or {},
-            })
-    except Exception as e:
-        logger.warning(f"Album search failed for {type(client).__name__}: {e}", exc_info=True)
-
-    try:
-        track_objs = client.search_tracks(query, limit=10)
-        for track in track_objs:
-            artist_name = ', '.join(track.artists) if track.artists else 'Unknown Artist'
-            tracks.append({
-                "id": track.id,
-                "name": track.name,
-                "artist": artist_name,
-                "album": track.album,
-                "duration_ms": track.duration_ms,
-                "image_url": track.image_url,
-                "release_date": track.release_date,
-                "external_urls": track.external_urls or {},
-            })
-    except Exception as e:
-        logger.warning(f"Track search failed for {type(client).__name__}: {e}", exc_info=True)
-
-    return {"artists": artists, "albums": albums, "tracks": tracks, "available": True}
+    return {"artists": results["artists"], "albums": results["albums"], "tracks": results["tracks"], "available": True}
 
 
 @app.route('/api/enhanced-search/source/<source_name>', methods=['POST'])
@@ -8766,51 +8875,20 @@ def enhanced_search_source(source_name):
 
         def generate():
             # Stream each search type as it completes
-            try:
-                artist_objs = client.search_artists(query, limit=10)
-                artists = []
-                for artist in artist_objs:
-                    artists.append({
-                        "id": artist.id, "name": artist.name,
-                        "image_url": artist.image_url,
-                        "external_urls": artist.external_urls or {},
-                    })
-                yield json.dumps({"type": "artists", "data": artists}) + "\n"
-            except Exception as e:
-                logger.debug(f"Artist search failed for {source_name}: {e}")
-                yield json.dumps({"type": "artists", "data": []}) + "\n"
-
-            try:
-                album_objs = client.search_albums(query, limit=10)
-                albums = []
-                for album in album_objs:
-                    artist_name = ', '.join(album.artists) if album.artists else 'Unknown Artist'
-                    albums.append({
-                        "id": album.id, "name": album.name, "artist": artist_name,
-                        "image_url": album.image_url, "release_date": album.release_date,
-                        "total_tracks": album.total_tracks, "album_type": album.album_type,
-                        "external_urls": album.external_urls or {},
-                    })
-                yield json.dumps({"type": "albums", "data": albums}) + "\n"
-            except Exception as e:
-                logger.warning(f"Album search failed for {source_name}: {e}")
-                yield json.dumps({"type": "albums", "data": []}) + "\n"
-
-            try:
-                track_objs = client.search_tracks(query, limit=10)
-                tracks = []
-                for track in track_objs:
-                    artist_name = ', '.join(track.artists) if track.artists else 'Unknown Artist'
-                    tracks.append({
-                        "id": track.id, "name": track.name, "artist": artist_name,
-                        "album": track.album, "duration_ms": track.duration_ms,
-                        "image_url": track.image_url, "release_date": track.release_date,
-                        "external_urls": track.external_urls or {},
-                    })
-                yield json.dumps({"type": "tracks", "data": tracks}) + "\n"
-            except Exception as e:
-                logger.warning(f"Track search failed for {source_name}: {e}")
-                yield json.dumps({"type": "tracks", "data": []}) + "\n"
+            with ThreadPoolExecutor(max_workers=3) as executor:
+                futures = {
+                    executor.submit(_search_metadata_source_kind, client, query, "artists", source_name): "artists",
+                    executor.submit(_search_metadata_source_kind, client, query, "albums", source_name): "albums",
+                    executor.submit(_search_metadata_source_kind, client, query, "tracks", source_name): "tracks",
+                }
+                for future in as_completed(futures):
+                    kind = futures[future]
+                    try:
+                        payload = future.result()
+                    except Exception as e:
+                        logger.warning(f"{kind.title()} search failed for {source_name}: {e}", exc_info=True)
+                        payload = []
+                    yield json.dumps({"type": kind, "data": payload}) + "\n"
 
             yield json.dumps({"type": "done"}) + "\n"
 
diff --git a/webui/static/script.js b/webui/static/script.js
index 84998221..974bbfce 100644
--- a/webui/static/script.js
+++ b/webui/static/script.js
@@ -8672,6 +8672,7 @@ function initializeSearchModeToggle() {
 
     async function performEnhancedSearch(query) {
         console.log('Enhanced search:', query);
+        const searchId = Date.now() + Math.random();
 
         // Show loading state with correct source name
         showDropdown();
@@ -8694,14 +8695,7 @@ function initializeSearchModeToggle() {
         _altSourceController = new AbortController();
 
         // Initialize multi-source state early so alternate fetches can write to it
-        _enhancedSearchData = { db_artists: [], primary_source: null, sources: {} };
-
-        // Fire ALL source fetches immediately in parallel with the primary endpoint.
-        // Don't guess which is primary — the main endpoint response will tell us.
-        // If an alternate duplicates the primary, it just overwrites with same data.
-        for (const srcName of ['spotify', 'itunes', 'deezer', 'discogs', 'hydrabase', 'youtube_videos', 'musicbrainz']) {
-            _fetchAlternateSource(srcName, query);
-        }
+        _enhancedSearchData = { db_artists: [], primary_source: null, sources: {}, searchId, query };
 
         try {
             const response = await fetch('/api/enhanced-search', {
@@ -8717,7 +8711,7 @@ function initializeSearchModeToggle() {
             console.log('Enhanced results:', data);
 
             // Store multi-source state
-            const primarySource = data.primary_source || data.metadata_source || 'spotify';
+            const primarySource = data.primary_source || data.metadata_source || 'deezer';
             _activeSearchSource = primarySource;
             _enhancedSearchData = _enhancedSearchData || {};
             _enhancedSearchData.db_artists = data.db_artists;
@@ -8747,6 +8741,10 @@ function initializeSearchModeToggle() {
                 resultsContainer.classList.remove('hidden');
             }
 
+            // Alternate sources now start after the primary response has landed.
+            // This avoids speculative fan-out for short or aborted searches.
+            _queueAlternateSourceFetches(data.alternate_sources || [], query, searchId);
+
         } catch (error) {
             if (error.name !== 'AbortError') {
                 console.error('Enhanced search error:', error);
@@ -8961,8 +8959,26 @@ function initializeSearchModeToggle() {
         }
     }
 
-    async function _fetchAlternateSource(sourceName, query) {
+    function _queueAlternateSourceFetches(alternateSources, query, searchId) {
+        if (!Array.isArray(alternateSources) || alternateSources.length === 0) return;
+
+        // Fetch metadata sources first, then YouTube last so it does not compete
+        // with the primary artist/album/track results for early attention.
+        const orderedSources = ['spotify', 'itunes', 'deezer', 'discogs', 'hydrabase', 'youtube_videos']
+            .filter(src => alternateSources.includes(src) && src !== _activeSearchSource);
+
+        orderedSources.forEach((src, index) => {
+            setTimeout(() => {
+                if (!_enhancedSearchData || _enhancedSearchData.searchId !== searchId) return;
+                _fetchAlternateSource(src, query, searchId);
+            }, index * 150);
+        });
+    }
+
+    async function _fetchAlternateSource(sourceName, query, searchId) {
         try {
+            if (!_enhancedSearchData || _enhancedSearchData.searchId !== searchId) return;
+
             const response = await fetch(`/api/enhanced-search/source/${sourceName}`, {
                 method: 'POST',
                 headers: { 'Content-Type': 'application/json' },
@@ -8970,9 +8986,9 @@ function initializeSearchModeToggle() {
                 signal: _altSourceController?.signal,
             });
             if (!response.ok) return;
+            if (!_enhancedSearchData || _enhancedSearchData.searchId !== searchId) return;
 
             // Stream NDJSON — render each search type (artists, albums, tracks) as it arrives
-            if (!_enhancedSearchData) return;
             if (!_enhancedSearchData.sources[sourceName]) {
                 const loadingSet = sourceName === 'youtube_videos' ? new Set(['videos']) : new Set(['artists', 'albums', 'tracks']);
                 _enhancedSearchData.sources[sourceName] = { artists: [], albums: [], tracks: [], videos: [], available: true, _loading: loadingSet };
@@ -8993,6 +9009,7 @@ function initializeSearchModeToggle() {
                     const line = buffer.slice(0, newlineIdx).trim();
                     buffer = buffer.slice(newlineIdx + 1);
                     if (!line) continue;
+                    if (!_enhancedSearchData || _enhancedSearchData.searchId !== searchId) return;
 
                     try {
                         const chunk = JSON.parse(line);
@@ -9016,7 +9033,7 @@ function initializeSearchModeToggle() {
             }
 
             // Final render
-            if (_enhancedSearchData.primary_source) {
+            if (_enhancedSearchData && _enhancedSearchData.searchId === searchId && _enhancedSearchData.primary_source) {
                 renderSourceTabs(_enhancedSearchData);
             }
         } catch (e) {
@@ -9033,16 +9050,25 @@ function initializeSearchModeToggle() {
         const sources = data.sources || {};
         const primary = data.primary_source || 'spotify';
 
-        // Build tab list: primary first, then alternates sorted alphabetically
+        // Build tab list: primary first, then alternates sorted alphabetically.
+        // Hide completed zero-result sources so the bar stays focused.
         const sourceNames = Object.keys(sources).filter(s => sources[s].available);
-        if (sourceNames.length <= 1) {
+        const visibleSources = sourceNames.filter(name => {
+            const src = sources[name] || {};
+            const count = name === 'youtube_videos'
+                ? (src.videos?.length || 0)
+                : (src.artists?.length || 0) + (src.albums?.length || 0) + (src.tracks?.length || 0);
+            const isLoading = !!(src._loading && src._loading.size > 0);
+            return isLoading || count > 0 || name === _activeSearchSource;
+        });
+        if (visibleSources.length <= 1) {
             tabBar.classList.add('hidden');
             tabBar.innerHTML = '';
             return;
         }
 
         // Primary tab first, then others
-        const ordered = [primary, ...sourceNames.filter(s => s !== primary).sort()];
+        const ordered = [primary, ...visibleSources.filter(s => s !== primary).sort()];
 
         tabBar.innerHTML = ordered.map(name => {
             const info = SOURCE_LABELS[name] || { text: name, tabClass: '' };
@@ -18169,10 +18195,26 @@ function _gsRenderTabs() {
     const el = document.getElementById('gsearch-tabs');
     if (!el) return;
     const sources = Object.keys(_gsState.sources);
-    if (sources.length < 2) { el.style.display = 'none'; return; }
-    const labels = { spotify: 'Spotify', itunes: 'Apple Music', deezer: 'Deezer', discogs: 'Discogs', hydrabase: 'Hydrabase', youtube_videos: 'Music Videos' };
+    const labels = {
+        spotify: 'Spotify',
+        itunes: 'Apple Music',
+        deezer: 'Deezer',
+        discogs: 'Discogs',
+        hydrabase: 'Hydrabase',
+        youtube_videos: 'Music Videos',
+        musicbrainz: 'MusicBrainz',
+    };
+    const visibleSources = sources.filter(s => {
+        const d = _gsState.sources[s] || {};
+        const count = s === 'youtube_videos'
+            ? (d.videos?.length || 0)
+            : (d.artists?.length || 0) + (d.albums?.length || 0) + (d.tracks?.length || 0);
+        const isLoading = !!(d._loading && d._loading.size > 0);
+        return isLoading || count > 0 || s === _gsState.activeSource;
+    });
+    if (visibleSources.length < 2) { el.style.display = 'none'; return; }
     el.style.display = 'flex';
-    el.innerHTML = sources.map(s => {
+    el.innerHTML = visibleSources.map(s => {
         const d = _gsState.sources[s];
         const c = s === 'youtube_videos'
             ? (d.videos?.length || 0)