feat(library): focused music/TV/movie landing screens + self-draining download queue

Library screens:
- Add dedicated music, TV, and movie landing pages (hero banner +
  horizontal carousels) backed by new music/tv/movies stores.
- Route tvshows libraries to /library/tv; surface rediscover ("haven't
  listened to in a while") albums via a new repository method across
  online/offline/hybrid repos plus the repository_get_rediscover_albums
  command.
- Add an A-Z jump bar for long alphabetically-sorted lists, with grid
  index anchors in LibraryGrid/LibraryListView/TrackList.
- Filter the "Podcasts" folder out of music library queries.

Downloads:
- Add a backend queue pump: enqueue_download / enqueue_video_downloads
  persist the resolved stream URL + target dir on each row (migration
  017), and the pump starts up to max_concurrent and drains the rest
  automatically as slots free, instead of the frontend silently dropping
  items past the concurrency limit. Album/series/season buttons now
  enqueue rather than calling start_download directly.

Other fixes:
- Hybrid search now returns instant cache results and pushes the merged
  cache+server union via a request-id-tagged search-event, so superseded
  queries can't clobber fresher results.
- URL-encode SearchTerm / genres / item types in online repo requests.
- Android: pause on audio-becoming-noisy (headphone/BT disconnect).

src-tauri/src/commands/download/mod.rs

@@ -2,8 +2,9 @@
 
 #[cfg(test)]
 use crate::utils::lock::MutexSafe;
+use std::path::PathBuf;
 use std::sync::{Arc, Mutex};
-use tauri::State;
+use tauri::{Manager, State};
 use log::{debug, error, info, warn};
 
 use crate::download::{DownloadInfo, DownloadManager};
@@ -818,9 +819,7 @@ pub async fn start_download(
     stream_url: String,
     target_dir: String,
 ) -> Result<(), String> {
-    use crate::download::{DownloadTask, DownloadWorker};
     use crate::download::events::DownloadEvent;
-    use std::path::PathBuf;
     use tauri::Emitter;
 
     debug!("start_download called for download_id: {}", download_id);
@@ -906,16 +905,27 @@ pub async fn start_download(
         }
     };
 
-    // Update status to downloading and save file_size if we got it
+    // Update status to downloading and save file_size if we got it.
+    // Also persist the resolved stream URL + target dir so the queue pump can
+    // restart/resume this download by itself if needed.
     let update_query = if let Some(size) = file_size_from_server {
         Query::with_params(
-            "UPDATE downloads SET status = 'downloading', started_at = CURRENT_TIMESTAMP, file_size = ? WHERE id = ?",
-            vec![QueryParam::Int64(size), QueryParam::Int64(download_id)],
+            "UPDATE downloads SET status = 'downloading', started_at = CURRENT_TIMESTAMP, file_size = ?, stream_url = ?, target_dir = ? WHERE id = ?",
+            vec![
+                QueryParam::Int64(size),
+                QueryParam::String(stream_url.clone()),
+                QueryParam::String(target_dir.clone()),
+                QueryParam::Int64(download_id),
+            ],
         )
     } else {
         Query::with_params(
-            "UPDATE downloads SET status = 'downloading', started_at = CURRENT_TIMESTAMP WHERE id = ?",
-            vec![QueryParam::Int64(download_id)],
+            "UPDATE downloads SET status = 'downloading', started_at = CURRENT_TIMESTAMP, stream_url = ?, target_dir = ? WHERE id = ?",
+            vec![
+                QueryParam::String(stream_url.clone()),
+                QueryParam::String(target_dir.clone()),
+                QueryParam::Int64(download_id),
+            ],
         )
     };
 
@@ -936,29 +946,304 @@ pub async fn start_download(
     // Build target path
     let target_path = PathBuf::from(&target_dir).join(&file_path);
 
-    // Create download task
-    let task = DownloadTask {
-        url: stream_url,
-        target_path: target_path.clone(),
-    };
-
-    // Start download in background
-    let app_clone = app.clone();
-    let item_id_clone = item_id.clone();
-
     // Get a clone of the active downloads Arc for unregistering later
     let active_downloads = {
         let manager = download_manager.0.lock().map_err(|e| e.to_string())?;
         manager.get_active_downloads()
     };
 
+    // Run the worker in the background; on completion/failure it frees the slot
+    // and pumps the next pending download.
+    spawn_download_worker(
+        app.clone(),
+        download_id,
+        item_id,
+        stream_url,
+        target_path,
+        active_downloads,
+    );
+
+    Ok(())
+}
+
+/// Enqueue a download with its resolved stream URL, then let the queue pump
+/// start it (or a higher-priority pending item) when a slot is free.
+///
+/// Unlike [`start_download`], this never errors when the concurrency limit is
+/// reached: the URL is persisted on the row and the pump will pick it up once a
+/// slot frees. This is the path bulk operations (album/series/season) use so
+/// every queued item eventually downloads without the frontend re-issuing it.
+#[tauri::command]
+#[specta::specta]
+pub async fn enqueue_download(
+    db: State<'_, DatabaseWrapper>,
+    download_manager: State<'_, DownloadManagerWrapper>,
+    app: tauri::AppHandle,
+    download_id: i64,
+    stream_url: String,
+    target_dir: String,
+) -> Result<(), String> {
+    let db_service = {
+        let database = db.0.lock().map_err(|e| e.to_string())?;
+        Arc::new(database.service())
+    };
+
+    // Persist the resolved URL/dir and mark the row pending so the pump can
+    // start it. We don't flip to 'downloading' here — the pump owns that.
+    let update_query = Query::with_params(
+        "UPDATE downloads SET status = 'pending', stream_url = ?, target_dir = ? WHERE id = ?",
+        vec![
+            QueryParam::String(stream_url),
+            QueryParam::String(target_dir),
+            QueryParam::Int64(download_id),
+        ],
+    );
+    db_service.execute(update_query).await.map_err(|e| e.to_string())?;
+
+    // Kick the pump: it will start as many pending downloads as there are slots.
+    let active_downloads = {
+        let manager = download_manager.0.lock().map_err(|e| e.to_string())?;
+        manager.get_active_downloads()
+    };
+    pump_download_queue(app, db_service, active_downloads).await;
+
+    Ok(())
+}
+
+/// Enqueue a batch of already-queued video downloads, resolving each one's
+/// transcode URL from the repository using the `quality_preset` stored on the
+/// row. Then let the pump start them subject to the concurrency limit.
+///
+/// This is the bulk video path (series/season): `download_series`/
+/// `download_season` insert the rows, then this resolves URLs and enqueues them
+/// so they actually start. Resolving server-side avoids round-tripping every
+/// episode URL through the frontend.
+#[tauri::command]
+#[specta::specta]
+pub async fn enqueue_video_downloads(
+    db: State<'_, DatabaseWrapper>,
+    download_manager: State<'_, DownloadManagerWrapper>,
+    repository: State<'_, crate::commands::repository::RepositoryManagerWrapper>,
+    app: tauri::AppHandle,
+    handle: String,
+    download_ids: Vec<i64>,
+    target_dir: String,
+) -> Result<(), String> {
+    use crate::repository::MediaRepository;
+
+    let repo = repository.0.get(&handle).ok_or("Repository not found")?;
+
+    let db_service = {
+        let database = db.0.lock().map_err(|e| e.to_string())?;
+        Arc::new(database.service())
+    };
+
+    for download_id in download_ids {
+        // Read the item + quality preset for this queued download.
+        let info_query = Query::with_params(
+            "SELECT item_id, COALESCE(quality_preset, 'original') FROM downloads WHERE id = ?",
+            vec![QueryParam::Int64(download_id)],
+        );
+        let (item_id, quality): (String, String) = match db_service
+            .query_one(info_query, |row| Ok((row.get(0)?, row.get(1)?)))
+            .await
+        {
+            Ok(row) => row,
+            Err(e) => {
+                warn!("[enqueue_video] Skipping download {}: {}", download_id, e);
+                continue;
+            }
+        };
+
+        // Build the transcode URL (pure URL builder, no server round-trip).
+        let stream_url = repo.as_ref().get_video_download_url(&item_id, &quality, None);
+
+        let update_query = Query::with_params(
+            "UPDATE downloads SET status = 'pending', stream_url = ?, target_dir = ? WHERE id = ?",
+            vec![
+                QueryParam::String(stream_url),
+                QueryParam::String(target_dir.clone()),
+                QueryParam::Int64(download_id),
+            ],
+        );
+        if let Err(e) = db_service.execute(update_query).await {
+            warn!("[enqueue_video] Failed to persist URL for download {}: {}", download_id, e);
+        }
+    }
+
+    // Pump once: starts up to max_concurrent, the rest drain as slots free.
+    let active_downloads = {
+        let manager = download_manager.0.lock().map_err(|e| e.to_string())?;
+        manager.get_active_downloads()
+    };
+    pump_download_queue(app, db_service, active_downloads).await;
+
+    Ok(())
+}
+
+/// Start as many pending downloads as there are free concurrency slots.
+///
+/// Picks the highest-priority `pending` rows that have a persisted `stream_url`
+/// (FIFO within a priority), registers each, flips it to `downloading`, and
+/// spawns a worker. Each spawned worker calls this again on completion/failure,
+/// so the queue drains itself without any frontend involvement.
+async fn pump_download_queue(
+    app: tauri::AppHandle,
+    db_service: Arc<crate::storage::db_service::RusqliteService>,
+    active_downloads: Arc<Mutex<std::collections::HashSet<i64>>>,
+) {
+    use crate::download::events::DownloadEvent;
+    use tauri::Emitter;
+
+    let max_concurrent = {
+        let manager = app.state::<DownloadManagerWrapper>();
+        let manager = match manager.0.lock() {
+            Ok(m) => m,
+            Err(e) => {
+                error!("[pump] Failed to lock download manager: {}", e);
+                return;
+            }
+        };
+        manager.max_concurrent()
+    };
+
+    loop {
+        // How many slots are free right now?
+        let free_slots = {
+            let active = match active_downloads.lock() {
+                Ok(a) => a,
+                Err(e) => {
+                    error!("[pump] Failed to lock active downloads: {}", e);
+                    return;
+                }
+            };
+            max_concurrent.saturating_sub(active.len())
+        };
+        if free_slots == 0 {
+            return;
+        }
+
+        // Find the next pending, startable download (has a stream URL). Exclude
+        // anything already registered as active to avoid double-starting.
+        let next_query = Query::with_params(
+            "SELECT id, item_id, file_path, stream_url, target_dir
+             FROM downloads
+             WHERE status = 'pending'
+               AND stream_url IS NOT NULL
+               AND target_dir IS NOT NULL
+             ORDER BY priority DESC, queued_at ASC",
+            vec![],
+        );
+
+        let candidates: Vec<(i64, String, String, String, String)> = match db_service
+            .query_many(next_query, |row| {
+                Ok((row.get(0)?, row.get(1)?, row.get(2)?, row.get(3)?, row.get(4)?))
+            })
+            .await
+        {
+            Ok(rows) => rows,
+            Err(e) => {
+                error!("[pump] Failed to query pending downloads: {}", e);
+                return;
+            }
+        };
+
+        // Pick the first candidate not already active.
+        let next = candidates.into_iter().find(|(id, _, _, _, _)| {
+            active_downloads
+                .lock()
+                .map(|active| !active.contains(id))
+                .unwrap_or(false)
+        });
+
+        let (download_id, item_id, file_path, stream_url, target_dir) = match next {
+            Some(n) => n,
+            None => return, // Nothing pending to start
+        };
+
+        // Register the slot. If registration fails (race: another pump filled
+        // the last slot), stop — we'll be re-pumped when a slot frees.
+        {
+            let manager = app.state::<DownloadManagerWrapper>();
+            let manager = match manager.0.lock() {
+                Ok(m) => m,
+                Err(e) => {
+                    error!("[pump] Failed to lock download manager: {}", e);
+                    return;
+                }
+            };
+            if !manager.register_download(download_id) {
+                return;
+            }
+            info!(
+                "[pump] Download {} started. Active downloads: {}/{}",
+                download_id,
+                manager.active_count(),
+                manager.max_concurrent()
+            );
+        }
+
+        // Mark as downloading and stamp started_at.
+        let update_query = Query::with_params(
+            "UPDATE downloads SET status = 'downloading', started_at = CURRENT_TIMESTAMP WHERE id = ?",
+            vec![QueryParam::Int64(download_id)],
+        );
+        if let Err(e) = db_service.execute(update_query).await {
+            error!("[pump] Failed to mark download {} downloading: {}", download_id, e);
+            if let Ok(mut a) = active_downloads.lock() {
+                a.remove(&download_id);
+            }
+            continue;
+        }
+
+        // Emit started event so the UI flips the row.
+        let _ = app.emit(
+            "download-event",
+            DownloadEvent::Started {
+                download_id,
+                item_id: item_id.clone(),
+            },
+        );
+
+        let target_path = PathBuf::from(&target_dir).join(&file_path);
+        spawn_download_worker(
+            app.clone(),
+            download_id,
+            item_id,
+            stream_url,
+            target_path,
+            active_downloads.clone(),
+        );
+    }
+}
+
+/// Spawn the background worker for one download. On completion or failure it
+/// unregisters the slot, emits the terminal event, and pumps the queue so the
+/// next pending download starts automatically.
+fn spawn_download_worker(
+    app: tauri::AppHandle,
+    download_id: i64,
+    item_id: String,
+    stream_url: String,
+    target_path: std::path::PathBuf,
+    active_downloads: Arc<Mutex<std::collections::HashSet<i64>>>,
+) {
+    use crate::download::{DownloadTask, DownloadWorker};
+    use crate::download::events::DownloadEvent;
+    use tauri::Emitter;
+
+    let task = DownloadTask {
+        url: stream_url,
+        target_path: target_path.clone(),
+    };
+
     tauri::async_runtime::spawn(async move {
         debug!("Download task started for download_id: {}", download_id);
         let worker = DownloadWorker::new();
 
         // Progress callback that emits events to the frontend
-        let progress_app = app_clone.clone();
-        let progress_item_id = item_id_clone.clone();
+        let progress_app = app.clone();
+        let progress_item_id = item_id.clone();
         let on_progress = move |bytes_downloaded: u64, total_bytes: Option<u64>| {
             let progress = total_bytes
                 .filter(|&t| t > 0)
@@ -975,54 +1260,55 @@ pub async fn start_download(
             let _ = progress_app.emit("download-event", event);
         };
 
-        match worker.download(&task, on_progress).await {
-            Ok(result) => {
-                info!("Download completed successfully: {} bytes", result.bytes_downloaded);
+        let result = worker.download(&task, on_progress).await;
 
-                // Unregister from download manager
-                if let Ok(mut active) = active_downloads.lock() {
-                    active.remove(&download_id);
-                    debug!("   Unregistered download {}. Active downloads: {}", download_id, active.len());
-                }
+        // Free the slot before pumping so the next download can take it.
+        if let Ok(mut active) = active_downloads.lock() {
+            active.remove(&download_id);
+            debug!("   Unregistered download {}. Active downloads: {}", download_id, active.len());
+        }
 
-                // Emit completed event - the frontend will handle state updates
+        match result {
+            Ok(res) => {
+                info!("Download completed successfully: {} bytes", res.bytes_downloaded);
                 let completed_event = DownloadEvent::Completed {
                     download_id,
-                    item_id: item_id_clone,
+                    item_id,
                     file_path: target_path.to_string_lossy().to_string(),
                 };
-                debug!("Emitting completed event: {:?}", completed_event);
-                debug!("   Serialized: {}", serde_json::to_string(&completed_event).unwrap_or_default());
-                match app_clone.emit("download-event", completed_event) {
+                match app.emit("download-event", completed_event) {
                     Ok(_) => debug!("   Completed event emitted successfully"),
                     Err(e) => error!("   Completed event emit failed: {:?}", e),
                 }
             }
             Err(e) => {
                 error!("Download failed: {:?}", e);
-
-                // Unregister from download manager
-                if let Ok(mut active) = active_downloads.lock() {
-                    active.remove(&download_id);
-                    debug!("   Unregistered failed download {}. Active downloads: {}", download_id, active.len());
-                }
-
-                // Emit failed event - the frontend will handle state updates
                 let failed_event = DownloadEvent::Failed {
                     download_id,
-                    item_id: item_id_clone.clone(),
+                    item_id,
                     error: e.to_string(),
                 };
-                debug!("Emitting failed event: {:?}", failed_event);
-                match app_clone.emit("download-event", failed_event) {
+                match app.emit("download-event", failed_event) {
                     Ok(_) => debug!("   Failed event emitted successfully"),
                     Err(e) => error!("   Failed event emit failed: {:?}", e),
                 }
             }
         }
-    });
 
-    Ok(())
+        // A slot just freed — start the next pending download (if any).
+        let db_service = {
+            let db = app.state::<DatabaseWrapper>();
+            let database = match db.0.lock() {
+                Ok(d) => d,
+                Err(e) => {
+                    error!("[pump] Failed to lock database after download {}: {}", download_id, e);
+                    return;
+                }
+            };
+            Arc::new(database.service())
+        };
+        pump_download_queue(app.clone(), db_service, active_downloads).await;
+    });
 }
 
 /// Delete a completed download

src-tauri/src/commands/repository.rs

@@ -7,8 +7,9 @@ use crate::utils::lock::MutexSafe;
 use std::collections::HashMap;
 use std::sync::{Arc, Mutex};
 
-use log::{debug, error, info};
-use tauri::State;
+use log::{debug, error, info, warn};
+use serde::{Deserialize, Serialize};
+use tauri::{AppHandle, Emitter, State};
 use uuid::Uuid;
 
 use crate::jellyfin::HttpClient;
@@ -259,6 +260,21 @@ pub async fn repository_get_resume_movies(
         .map_err(|e| format!("{:?}", e))
 }
 
+/// Get albums the user hasn't listened to recently ("rediscover")
+#[tauri::command]
+#[specta::specta]
+pub async fn repository_get_rediscover_albums(
+    manager: State<'_, RepositoryManagerWrapper>,
+    handle: String,
+    parent_id: Option<String>,
+    limit: Option<usize>,
+) -> Result<Vec<MediaItem>, String> {
+    let repo = manager.0.get(&handle).ok_or("Repository not found")?;
+    repo.as_ref().get_rediscover_albums(parent_id.as_deref(), limit)
+        .await
+        .map_err(|e| format!("{:?}", e))
+}
+
 /// Get genres for a library
 #[tauri::command]
 #[specta::specta]
@@ -273,19 +289,73 @@ pub async fn repository_get_genres(
         .map_err(|e| format!("{:?}", e))
 }
 
+/// Tauri event name carrying the merged (cache + server) search results.
+pub const SEARCH_EVENT_NAME: &str = "search-event";
+
+/// Payload for the deferred, merged search results pushed to the frontend.
+///
+/// `request_id` matches the value the frontend passed to `repository_search`,
+/// letting it discard updates from queries that have since been superseded.
+#[derive(specta::Type, Debug, Clone, Serialize, Deserialize)]
+#[serde(rename_all = "camelCase")]
+pub struct SearchUpdateEvent {
+    pub request_id: u32,
+    pub result: SearchResult,
+}
+
 /// Search for items
 #[tauri::command]
 #[specta::specta]
 pub async fn repository_search(
+    app: AppHandle,
     manager: State<'_, RepositoryManagerWrapper>,
     handle: String,
     query: String,
     options: Option<SearchOptions>,
+    request_id: u32,
 ) -> Result<SearchResult, String> {
     let repo = manager.0.get(&handle).ok_or("Repository not found")?;
-    repo.as_ref().search(&query, options)
+
+    // Phase 1: instant local results from the cache (downloaded content) so the
+    // UI can render immediately while the server is still being queried.
+    let cache_result = repo
+        .search_cache_only(&query, options.clone())
         .await
-        .map_err(|e| format!("{:?}", e))
+        .unwrap_or_else(|e| {
+            debug!("[Search] Cache search miss/timeout: {:?}", e);
+            SearchResult {
+                items: Vec::new(),
+                total_record_count: 0,
+            }
+        });
+
+    // Phase 2: query the live server in the background, merge with the cache,
+    // and push the union to the frontend via a `search-event`. Tagged with
+    // `request_id` so the frontend can discard results from superseded queries.
+    let repo_bg = repo.clone();
+    let cache_for_merge = cache_result.clone();
+    tauri::async_runtime::spawn(async move {
+        match repo_bg.search_server_only(&query, options).await {
+            Ok(server_result) => {
+                let merged =
+                    HybridRepository::merge_search_results(cache_for_merge, server_result);
+                let event = SearchUpdateEvent {
+                    request_id,
+                    result: merged,
+                };
+                if let Err(e) = app.emit(SEARCH_EVENT_NAME, &event) {
+                    error!("[Search] Failed to emit search update: {}", e);
+                }
+            }
+            Err(e) => {
+                // Server failed — the cache results are already on screen, so
+                // just log. (Offline / unreachable server falls here.)
+                warn!("[Search] Server search failed, keeping cache results: {:?}", e);
+            }
+        }
+    });
+
+    Ok(cache_result)
 }
 
 /// Get playback info for an item

src-tauri/src/lib.rs

@@ -28,7 +28,7 @@ use commands::{
     get_download_storage_stats, get_downloads, get_download_manager_stats, set_max_concurrent_downloads,
     get_smart_cache_stats, update_smart_cache_config, get_smart_cache_config, get_album_recommendations,
     get_album_affinity_status,
-    mark_download_completed, mark_download_failed, start_download,
+    mark_download_completed, mark_download_failed, start_download, enqueue_download, enqueue_video_downloads,
     pin_item, unpin_item, is_item_pinned,
     offline_get_items, offline_is_available, offline_search, pause_download, resume_download,
     player_cycle_repeat, player_get_audio_settings, player_get_queue, player_get_status,
@@ -96,6 +96,7 @@ use commands::{
     repository_create, repository_destroy, repository_get_libraries, repository_get_items,
     repository_get_item, repository_get_latest_items, repository_get_resume_items,
     repository_get_next_up_episodes, repository_get_recently_played_audio, repository_get_resume_movies,
+    repository_get_rediscover_albums,
     repository_get_genres, repository_search, repository_get_playback_info,
     repository_get_video_stream_url, repository_get_audio_stream_url,
     repository_report_playback_start, repository_report_playback_progress, repository_report_playback_stopped,
@@ -493,6 +494,8 @@ fn specta_builder() -> Builder<tauri::Wry> {
             mark_download_completed,
             mark_download_failed,
             start_download,
+            enqueue_download,
+            enqueue_video_downloads,
             get_download_manager_stats,
             set_max_concurrent_downloads,
             get_smart_cache_stats,
@@ -552,6 +555,7 @@ fn specta_builder() -> Builder<tauri::Wry> {
             repository_get_next_up_episodes,
             repository_get_recently_played_audio,
             repository_get_resume_movies,
+            repository_get_rediscover_albums,
             repository_get_genres,
             repository_search,
             repository_get_playback_info,

src-tauri/src/repository/hybrid.rs

@@ -57,6 +57,81 @@ impl HybridRepository {
         self.online.get_video_stream_url(item_id, media_source_id, start_time_seconds, audio_stream_index).await
     }
 
+    /// Search only the local SQLite cache (downloaded content).
+    ///
+    /// Fast (100ms timeout) — used to render instant results before the server
+    /// responds. Returns an empty result rather than erroring on timeout so the
+    /// caller can still fall through to the server.
+    pub async fn search_cache_only(
+        &self,
+        query: &str,
+        options: Option<SearchOptions>,
+    ) -> Result<SearchResult, RepoError> {
+        let offline = Arc::clone(&self.offline);
+        let query = query.to_string();
+        self.cache_with_timeout(async move { offline.search(&query, options).await })
+            .await
+    }
+
+    /// Search only the live Jellyfin server (full library).
+    pub async fn search_server_only(
+        &self,
+        query: &str,
+        options: Option<SearchOptions>,
+    ) -> Result<SearchResult, RepoError> {
+        self.online.search(query, options).await
+    }
+
+    /// Merge cache and server search results into a single de-duplicated list.
+    ///
+    /// Ordering: local (cached/downloaded) items first, then server-only items
+    /// appended. On a duplicate `id`, the server's item wins (fresher, more
+    /// complete metadata) but keeps the local item's earlier position.
+    pub fn merge_search_results(cache: SearchResult, server: SearchResult) -> SearchResult {
+        use std::collections::HashMap;
+
+        // Index server items by id so we can (a) override duplicates with the
+        // server's metadata and (b) know which server items are brand new.
+        let mut server_by_id: HashMap<String, MediaItem> = HashMap::new();
+        let mut server_order: Vec<String> = Vec::with_capacity(server.items.len());
+        for item in server.items {
+            if !server_by_id.contains_key(&item.id) {
+                server_order.push(item.id.clone());
+            }
+            server_by_id.insert(item.id.clone(), item);
+        }
+
+        let mut items: Vec<MediaItem> = Vec::new();
+        let mut seen: std::collections::HashSet<String> = std::collections::HashSet::new();
+
+        // Local items first, in their original order. If the server also
+        // returned this item, take the server's copy (newer metadata).
+        for local in cache.items {
+            if !seen.insert(local.id.clone()) {
+                continue;
+            }
+            match server_by_id.remove(&local.id) {
+                Some(server_item) => items.push(server_item),
+                None => items.push(local),
+            }
+        }
+
+        // Then append server-only items, preserving the server's order.
+        for id in server_order {
+            if let Some(server_item) = server_by_id.remove(&id) {
+                if seen.insert(id) {
+                    items.push(server_item);
+                }
+            }
+        }
+
+        let total_record_count = items.len();
+        SearchResult {
+            items,
+            total_record_count,
+        }
+    }
+
     /// Cache-first query: try cache, fall back to server on miss.
     ///
     /// 1. Check cache (100ms timeout applied by caller via cache_with_timeout)
@@ -274,6 +349,27 @@ impl MediaRepository for HybridRepository {
         self.parallel_race(cache_future, server_future).await
     }
 
+    async fn get_rediscover_albums(
+        &self,
+        parent_id: Option<&str>,
+        limit: Option<usize>,
+    ) -> Result<Vec<MediaItem>, RepoError> {
+        let offline = Arc::clone(&self.offline);
+        let online = Arc::clone(&self.online);
+        let parent_id_owned = parent_id.map(|s| s.to_string());
+        let parent_id_clone = parent_id_owned.clone();
+
+        let cache_future = self.cache_with_timeout(async move {
+            offline.get_rediscover_albums(parent_id_owned.as_deref(), limit).await
+        });
+
+        let server_future = async move {
+            online.get_rediscover_albums(parent_id_clone.as_deref(), limit).await
+        };
+
+        self.parallel_race(cache_future, server_future).await
+    }
+
     async fn get_genres(&self, parent_id: Option<&str>) -> Result<Vec<Genre>, RepoError> {
         let offline = Arc::clone(&self.offline);
         let online = Arc::clone(&self.online);
@@ -602,6 +698,14 @@ mod tests {
             unimplemented!()
         }
 
+        async fn get_rediscover_albums(
+            &self,
+            _parent_id: Option<&str>,
+            _limit: Option<usize>,
+        ) -> Result<Vec<MediaItem>, RepoError> {
+            unimplemented!()
+        }
+
         async fn get_genres(&self, _parent_id: Option<&str>) -> Result<Vec<Genre>, RepoError> {
             unimplemented!()
         }
@@ -759,6 +863,14 @@ mod tests {
             unimplemented!()
         }
 
+        async fn get_rediscover_albums(
+            &self,
+            _parent_id: Option<&str>,
+            _limit: Option<usize>,
+        ) -> Result<Vec<MediaItem>, RepoError> {
+            unimplemented!()
+        }
+
         async fn get_genres(&self, _parent_id: Option<&str>) -> Result<Vec<Genre>, RepoError> {
             unimplemented!()
         }
@@ -1058,4 +1170,75 @@ mod tests {
         };
         assert!(result_with_items.has_content(), "Result with items should have content");
     }
+
+    #[test]
+    fn test_merge_search_local_first_then_server_appended() {
+        let cache = SearchResult {
+            items: vec![
+                create_test_item("a", "Cached A"),
+                create_test_item("b", "Cached B"),
+            ],
+            total_record_count: 2,
+        };
+        let server = SearchResult {
+            items: vec![
+                create_test_item("c", "Server C"),
+                create_test_item("d", "Server D"),
+            ],
+            total_record_count: 2,
+        };
+
+        let merged = HybridRepository::merge_search_results(cache, server);
+
+        // Local items first (in order), then server-only items appended.
+        let ids: Vec<&str> = merged.items.iter().map(|i| i.id.as_str()).collect();
+        assert_eq!(ids, vec!["a", "b", "c", "d"]);
+        assert_eq!(merged.total_record_count, 4);
+    }
+
+    #[test]
+    fn test_merge_search_dedupes_with_server_winning() {
+        // "b" appears in both. Server metadata should win, but the item keeps
+        // its earlier (local) position and is not duplicated.
+        let cache = SearchResult {
+            items: vec![
+                create_test_item("a", "Cached A"),
+                create_test_item("b", "Cached B"),
+            ],
+            total_record_count: 2,
+        };
+        let server = SearchResult {
+            items: vec![
+                create_test_item("b", "Server B (fresher)"),
+                create_test_item("c", "Server C"),
+            ],
+            total_record_count: 2,
+        };
+
+        let merged = HybridRepository::merge_search_results(cache, server);
+
+        let ids: Vec<&str> = merged.items.iter().map(|i| i.id.as_str()).collect();
+        assert_eq!(ids, vec!["a", "b", "c"], "no duplicate, local position kept");
+
+        let b = merged.items.iter().find(|i| i.id == "b").unwrap();
+        assert_eq!(b.name, "Server B (fresher)", "server metadata wins on conflict");
+        assert_eq!(merged.total_record_count, 3);
+    }
+
+    #[test]
+    fn test_merge_search_handles_empty_sides() {
+        let only_server = HybridRepository::merge_search_results(
+            SearchResult { items: vec![], total_record_count: 0 },
+            SearchResult { items: vec![create_test_item("x", "X")], total_record_count: 1 },
+        );
+        assert_eq!(only_server.items.len(), 1);
+        assert_eq!(only_server.items[0].id, "x");
+
+        let only_cache = HybridRepository::merge_search_results(
+            SearchResult { items: vec![create_test_item("y", "Y")], total_record_count: 1 },
+            SearchResult { items: vec![], total_record_count: 0 },
+        );
+        assert_eq!(only_cache.items.len(), 1);
+        assert_eq!(only_cache.items[0].id, "y");
+    }
 }

src-tauri/src/repository/mod.rs

@@ -79,6 +79,15 @@ pub trait MediaRepository: Send + Sync {
         limit: Option<usize>,
     ) -> Result<Vec<MediaItem>, RepoError>;
 
+    /// Get albums the user has played, but not recently ("rediscover" / haven't
+    /// listened to in a while). Returns albums sorted by least-recently played
+    /// first, optionally restricted to a parent library.
+    async fn get_rediscover_albums(
+        &self,
+        parent_id: Option<&str>,
+        limit: Option<usize>,
+    ) -> Result<Vec<MediaItem>, RepoError>;
+
     /// Get resume movies
     async fn get_resume_movies(&self, limit: Option<usize>) -> Result<Vec<MediaItem>, RepoError>;
 

src-tauri/src/repository/offline.rs

@@ -778,6 +778,17 @@ impl MediaRepository for OfflineRepository {
         Ok(items)
     }
 
+    async fn get_rediscover_albums(
+        &self,
+        _parent_id: Option<&str>,
+        _limit: Option<usize>,
+    ) -> Result<Vec<MediaItem>, RepoError> {
+        // "Rediscover" is a discovery feature over the full server library.
+        // Offline only holds downloaded items, so there is nothing meaningful
+        // to surface here; the hybrid repo serves this from the server instead.
+        Ok(Vec::new())
+    }
+
     async fn get_resume_movies(&self, limit: Option<usize>) -> Result<Vec<MediaItem>, RepoError> {
         let limit_val = limit.unwrap_or(12);
 

src-tauri/src/repository/online.rs

@@ -568,6 +568,16 @@ impl MediaRepository for OnlineRepository {
             if let Some(recursive) = opts.recursive {
                 endpoint.push_str(&format!("&Recursive={}", recursive));
             }
+            if let Some(genres) = opts.genres {
+                if !genres.is_empty() {
+                    // Genre names may contain spaces/ampersands, so percent-encode each.
+                    let encoded: Vec<String> = genres
+                        .iter()
+                        .map(|g| urlencoding::encode(g).into_owned())
+                        .collect();
+                    endpoint.push_str(&format!("&Genres={}", encoded.join("|")));
+                }
+            }
         }
 
         // Request image fields for list views (People only needed in get_item detail view)
@@ -759,6 +769,32 @@ impl MediaRepository for OnlineRepository {
         Ok(final_result)
     }
 
+    async fn get_rediscover_albums(
+        &self,
+        parent_id: Option<&str>,
+        limit: Option<usize>,
+    ) -> Result<Vec<MediaItem>, RepoError> {
+        let limit_val = limit.unwrap_or(12);
+        // Ask Jellyfin for played albums sorted by least-recently played first.
+        // Filters=IsPlayed keeps only albums the user has actually listened to,
+        // and SortBy=DatePlayed ascending surfaces the ones they've neglected.
+        let mut endpoint = format!(
+            "/Users/{}/Items?SortBy=DatePlayed&SortOrder=Ascending&IncludeItemTypes=MusicAlbum&Limit={}&Recursive=true&Filters=IsPlayed&Fields=BackdropImageTags,ParentBackdropImageTags",
+            self.user_id, limit_val
+        );
+
+        if let Some(pid) = parent_id {
+            endpoint.push_str(&format!("&ParentId={}", pid));
+        }
+
+        let response: ItemsResponse = self.get_json(&endpoint).await?;
+        Ok(response
+            .items
+            .into_iter()
+            .map(|item| item.to_media_item(self.user_id.clone()))
+            .collect())
+    }
+
     async fn get_resume_movies(&self, limit: Option<usize>) -> Result<Vec<MediaItem>, RepoError> {
         let limit_str = limit.unwrap_or(16);
         let endpoint = format!(
@@ -811,14 +847,24 @@ impl MediaRepository for OnlineRepository {
         options: Option<SearchOptions>,
     ) -> Result<SearchResult, RepoError> {
         let limit = options.as_ref().and_then(|o| o.limit).unwrap_or(50);
+        // SearchTerm is arbitrary user input and must be percent-encoded so that
+        // spaces, ampersands, etc. don't corrupt the query string (a multi-word
+        // search like "Star Wars" would otherwise produce a malformed URL).
         let mut endpoint = format!(
             "/Users/{}/Items?SearchTerm={}&Limit={}&Recursive=true",
-            self.user_id, query, limit
+            self.user_id,
+            urlencoding::encode(query),
+            limit
         );
 
         if let Some(opts) = options {
             if let Some(types) = opts.include_item_types {
-                endpoint.push_str(&format!("&IncludeItemTypes={}", types.join(",")));
+                let encoded_types = types
+                    .iter()
+                    .map(|t| urlencoding::encode(t).into_owned())
+                    .collect::<Vec<_>>()
+                    .join(",");
+                endpoint.push_str(&format!("&IncludeItemTypes={}", encoded_types));
             }
         }
 
@@ -1768,4 +1814,13 @@ mod tests {
         assert_eq!(response.items[0].id, "item1");
         assert_eq!(response.items[1].id, "item2");
     }
+
+    #[test]
+    fn test_search_term_is_url_encoded() {
+        // A multi-word query (and one with a reserved character) must be
+        // percent-encoded before being placed in the SearchTerm query param,
+        // otherwise the request URL is malformed and search returns nothing.
+        assert_eq!(urlencoding::encode("Star Wars"), "Star%20Wars");
+        assert_eq!(urlencoding::encode("Tom & Jerry"), "Tom%20%26%20Jerry");
+    }
 }

src-tauri/src/storage/schema.rs

@@ -21,6 +21,7 @@ pub const MIGRATIONS: &[(&str, &str)] = &[
     ("014_series_audio_preferences", MIGRATION_014),
     ("015_device_id", MIGRATION_015),
     ("016_autoplay_max_episodes", MIGRATION_016),
+    ("017_downloads_resume_url", MIGRATION_017),
 ];
 
 /// Initial schema migration
@@ -667,3 +668,15 @@ const MIGRATION_016: &str = r#"
 ALTER TABLE user_player_settings
 ADD COLUMN autoplay_max_episodes INTEGER DEFAULT 0;
 "#;
+
+/// Migration to persist the resolved stream URL and target directory on each
+/// download row. This lets the backend queue pump start a pending download by
+/// itself (replaying the stored URL) once a concurrency slot frees up, instead
+/// of relying on the frontend to re-issue every queued item.
+const MIGRATION_017: &str = r#"
+-- Resolved download source URL and on-disk target directory, captured when the
+-- download is enqueued. Nullable: pre-existing rows and rows enqueued without a
+-- URL simply won't be auto-started by the pump.
+ALTER TABLE downloads ADD COLUMN stream_url TEXT;
+ALTER TABLE downloads ADD COLUMN target_dir TEXT;
+"#;