perf(api,indexer): optimiser pages, thumbnails, watcher et robustesse fd

- Pages: mode Original (zero-transcoding), ETag/304, cache index CBZ,
  préfetch next 2 pages, filtre Triangle par défaut
- Thumbnails: DCT scaling JPEG via jpeg-decoder (decode 7x plus rapide),
  img.thumbnail() pour resize, support format Original, fix JPEG RGBA8
- API fallback thumbnail: OutputFormat::Original + DCT scaling au lieu
  de WebP full-decode, retour (bytes, content_type) dynamique
- Watcher: remplacement notify par poll léger sans inotify/fd,
  skip poll quand job actif, snapshots en mémoire
- Jobs: mutex exclusif corrigé (tous statuts actifs, tous types exclusifs)
- Robustesse: suppression fs::canonicalize (problèmes fd Docker),
  list_folders avec erreurs explicites, has_children default true
- Backoffice: FormRow items-start pour alignement inputs avec helper text,
  labels settings clarifiés

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Cargo.lock

@@ -61,6 +61,7 @@ dependencies = [
  "chrono",
  "futures",
  "image",
+ "jpeg-decoder",
  "lru",
  "parsers",
  "rand 0.8.5",
@@ -215,12 +216,6 @@ version = "1.8.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "2af50177e190e07a26ab74f8b1efbfe2ef87da2116221318cb1c2e82baf7de06"
 
-[[package]]
-name = "bitflags"
-version = "1.3.2"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "bef38d45163c2f1dde094a7dfd33ccf595c92905c8f8f4fdc18d06fb1037718a"
-
 [[package]]
 name = "bitflags"
 version = "2.11.0"
@@ -634,15 +629,6 @@ dependencies = [
  "percent-encoding",
 ]
 
-[[package]]
-name = "fsevent-sys"
-version = "4.1.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "76ee7a02da4d231650c7cea31349b889be2f45ddb3ef3032d2ec8185f6313fd2"
-dependencies = [
- "libc",
-]
-
 [[package]]
 name = "futures"
 version = "0.3.32"
@@ -1143,7 +1129,7 @@ dependencies = [
  "chrono",
  "futures",
  "image",
- "notify",
+ "jpeg-decoder",
  "num_cpus",
  "parsers",
  "rayon",
@@ -1173,26 +1159,6 @@ dependencies = [
  "serde_core",
 ]
 
-[[package]]
-name = "inotify"
-version = "0.11.1"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "bd5b3eaf1a28b758ac0faa5a4254e8ab2705605496f1b1f3fbbc3988ad73d199"
-dependencies = [
- "bitflags 2.11.0",
- "inotify-sys",
- "libc",
-]
-
-[[package]]
-name = "inotify-sys"
-version = "0.1.5"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e05c02b5e89bff3b946cedeca278abc628fe811e604f027c45a8aa3cf793d0eb"
-dependencies = [
- "libc",
-]
-
 [[package]]
 name = "inout"
 version = "0.1.4"
@@ -1285,6 +1251,15 @@ dependencies = [
  "libc",
 ]
 
+[[package]]
+name = "jpeg-decoder"
+version = "0.3.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "00810f1d8b74be64b13dbf3db89ac67740615d6c891f0e7b6179326533011a07"
+dependencies = [
+ "rayon",
+]
+
 [[package]]
 name = "js-sys"
 version = "0.3.91"
@@ -1295,26 +1270,6 @@ dependencies = [
  "wasm-bindgen",
 ]
 
-[[package]]
-name = "kqueue"
-version = "1.1.1"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "eac30106d7dce88daf4a3fcb4879ea939476d5074a9b7ddd0fb97fa4bed5596a"
-dependencies = [
- "kqueue-sys",
- "libc",
-]
-
-[[package]]
-name = "kqueue-sys"
-version = "1.0.4"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ed9625ffda8729b85e45cf04090035ac368927b8cebc34898e7c120f52e4838b"
-dependencies = [
- "bitflags 1.3.2",
- "libc",
-]
-
 [[package]]
 name = "lazy_static"
 version = "1.5.0"
@@ -1358,7 +1313,7 @@ version = "0.1.14"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "1744e39d1d6a9948f4f388969627434e31128196de472883b39f148769bfe30a"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "libc",
  "plain",
  "redox_syscall 0.7.3",
@@ -1412,7 +1367,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "f560f57dfb9142a02d673e137622fd515d4231e51feb8b4af28d92647d83f35b"
 dependencies = [
  "aes",
- "bitflags 2.11.0",
+ "bitflags",
  "cbc",
  "chrono",
  "ecb",
@@ -1522,7 +1477,6 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "a69bcab0ad47271a0234d9422b131806bf3968021e5dc9328caf2d4cd58557fc"
 dependencies = [
  "libc",
- "log",
  "wasi",
  "windows-sys 0.61.2",
 ]
@@ -1563,33 +1517,6 @@ dependencies = [
  "nom",
 ]
 
-[[package]]
-name = "notify"
-version = "8.2.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4d3d07927151ff8575b7087f245456e549fea62edf0ec4e565a5ee50c8402bc3"
-dependencies = [
- "bitflags 2.11.0",
- "fsevent-sys",
- "inotify",
- "kqueue",
- "libc",
- "log",
- "mio",
- "notify-types",
- "walkdir",
- "windows-sys 0.60.2",
-]
-
-[[package]]
-name = "notify-types"
-version = "2.1.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "42b8cfee0e339a0337359f3c88165702ac6e600dc01c0cc9579a92d62b08477a"
-dependencies = [
- "bitflags 2.11.0",
-]
-
 [[package]]
 name = "nu-ansi-term"
 version = "0.50.3"
@@ -1728,7 +1655,7 @@ version = "0.8.37"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "6553f6604a52b3203db7b4e9d51eb4dd193cf455af9e56d40cab6575b547b679"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "bytemuck",
  "bytes",
  "chrono",
@@ -1820,7 +1747,7 @@ version = "0.18.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "60769b8b31b2a9f263dae2776c37b1b28ae246943cf719eb6946a1db05128a61"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "crc32fast",
  "fdeflate",
  "flate2",
@@ -2088,7 +2015,7 @@ version = "0.5.18"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "ed2bf2547551a7053d6fdfafda3f938979645c44812fbfcda098faae3f1a362d"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
 ]
 
 [[package]]
@@ -2097,7 +2024,7 @@ version = "0.7.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "6ce70a74e890531977d37e532c34d45e9055d2409ed08ddba14529471ed0be16"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
 ]
 
 [[package]]
@@ -2579,7 +2506,7 @@ checksum = "aa003f0038df784eb8fecbbac13affe3da23b45194bd57dba231c8f48199c526"
 dependencies = [
  "atoi",
  "base64",
- "bitflags 2.11.0",
+ "bitflags",
  "byteorder",
  "bytes",
  "chrono",
@@ -2623,7 +2550,7 @@ checksum = "db58fcd5a53cf07c184b154801ff91347e4c30d17a3562a635ff028ad5deda46"
 dependencies = [
  "atoi",
  "base64",
- "bitflags 2.11.0",
+ "bitflags",
  "byteorder",
  "chrono",
  "crc",
@@ -2921,7 +2848,7 @@ version = "0.6.8"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "d4e6559d53cc268e5031cd8429d05415bc4cb4aefc4aa5d6cc35fbf5b924a1f8"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "bytes",
  "futures-util",
  "http",
@@ -3076,7 +3003,7 @@ version = "0.5.8"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "92ec61343a630d2b50d13216dea5125e157d3fc180a7d3f447d22fe146b648fc"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "regex",
  "unrar_sys",
  "widestring",
@@ -3341,7 +3268,7 @@ version = "0.244.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "47b807c72e1bac69382b3a6fb3dbe8ea4c0ed87ff5629b8685ae6b9a611028fe"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "hashbrown 0.15.5",
  "indexmap",
  "semver",
@@ -3787,7 +3714,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "9d66ea20e9553b30172b5e831994e35fbde2d165325bec84fc43dbf6f4eb9cb2"
 dependencies = [
  "anyhow",
- "bitflags 2.11.0",
+ "bitflags",
  "indexmap",
  "log",
  "serde",

Cargo.toml

@@ -19,6 +19,7 @@ axum = "0.7"
 base64 = "0.22"
 chrono = { version = "0.4", features = ["serde"] }
 image = { version = "0.25", default-features = false, features = ["jpeg", "png", "webp"] }
+jpeg-decoder = "0.3"
 lru = "0.12"
 rayon = "1.10"
 reqwest = { version = "0.12", default-features = false, features = ["json", "rustls-tls"] }

apps/api/Cargo.toml

@@ -13,6 +13,7 @@ async-stream = "0.3"
 chrono.workspace = true
 futures = "0.3"
 image.workspace = true
+jpeg-decoder.workspace = true
 lru.workspace = true
 stripstream-core = { path = "../../crates/core" }
 parsers = { path = "../../crates/parsers" }

apps/api/src/books.rs

@@ -584,6 +584,17 @@ use axum::{
     response::IntoResponse,
 };
 
+/// Detect content type from thumbnail file extension.
+fn detect_thumbnail_content_type(path: &str) -> &'static str {
+    if path.ends_with(".jpg") || path.ends_with(".jpeg") {
+        "image/jpeg"
+    } else if path.ends_with(".png") {
+        "image/png"
+    } else {
+        "image/webp"
+    }
+}
+
 /// Get book thumbnail image
 #[utoipa::path(
     get,
@@ -612,9 +623,12 @@ pub async fn get_thumbnail(
     let row = row.ok_or_else(|| ApiError::not_found("book not found"))?;
     let thumbnail_path: Option<String> = row.get("thumbnail_path");
 
-    let data = if let Some(ref path) = thumbnail_path {
+    let (data, content_type) = if let Some(ref path) = thumbnail_path {
         match std::fs::read(path) {
-            Ok(bytes) => bytes,
+            Ok(bytes) => {
+                let ct = detect_thumbnail_content_type(path);
+                (bytes, ct)
+            }
             Err(_) => {
                 // File missing on disk (e.g. different mount in dev) — fall back to live render
                 crate::pages::render_book_page_1(&state, book_id, 300, 80).await?
@@ -626,7 +640,7 @@ pub async fn get_thumbnail(
     };
 
     let mut headers = HeaderMap::new();
-    headers.insert(header::CONTENT_TYPE, HeaderValue::from_static("image/webp"));
+    headers.insert(header::CONTENT_TYPE, HeaderValue::from_static(content_type));
     headers.insert(
         header::CACHE_CONTROL,
         HeaderValue::from_static("public, max-age=31536000, immutable"),

apps/api/src/index_jobs.rs

@@ -246,9 +246,9 @@ pub async fn list_folders(
         base_path.to_path_buf()
     };
     
-    // Ensure the path is within the libraries root
+    // Ensure the path is within the libraries root (avoid canonicalize — burns fd on Docker mounts)
-    let canonical_target = target_path.canonicalize().unwrap_or(target_path.clone());
+    let canonical_target = target_path.clone();
-    let canonical_base = base_path.canonicalize().unwrap_or(base_path.to_path_buf());
+    let canonical_base = base_path.to_path_buf();
     
     if !canonical_target.starts_with(&canonical_base) {
         return Err(ApiError::bad_request("Path is outside libraries root"));
@@ -263,19 +263,31 @@ pub async fn list_folders(
         0
     };
 
-    if let Ok(entries) = std::fs::read_dir(&canonical_target) {
+    let entries = std::fs::read_dir(&canonical_target)
-        for entry in entries.flatten() {
+        .map_err(|e| ApiError::internal(format!("cannot read directory {}: {}", canonical_target.display(), e)))?;
-            if entry.file_type().map(|ft| ft.is_dir()).unwrap_or(false) {
+
+    for entry in entries {
+        let entry = match entry {
+            Ok(e) => e,
+            Err(e) => {
+                tracing::warn!("[FOLDERS] entry error in {}: {}", canonical_target.display(), e);
+                continue;
+            }
+        };
+        let is_dir = match entry.file_type() {
+            Ok(ft) => ft.is_dir(),
+            Err(e) => {
+                tracing::warn!("[FOLDERS] cannot stat {}: {}", entry.path().display(), e);
+                continue;
+            }
+        };
+        if is_dir {
                 let name = entry.file_name().to_string_lossy().to_string();
 
-                // Check if this folder has children
+                // Check if this folder has children (best-effort, default to true on error)
-                let has_children = if let Ok(sub_entries) = std::fs::read_dir(entry.path()) {
+                let has_children = std::fs::read_dir(entry.path())
-                    sub_entries.flatten().any(|e| {
+                    .map(|sub| sub.flatten().any(|e| e.file_type().map(|ft| ft.is_dir()).unwrap_or(false)))
-                        e.file_type().map(|ft| ft.is_dir()).unwrap_or(false)
+                    .unwrap_or(true);
-                    })
-                } else {
-                    false
-                };
                 
                 // Calculate the full path relative to libraries root
                 let full_path = if let Ok(relative) = entry.path().strip_prefix(&canonical_base) {
@@ -292,7 +304,6 @@ pub async fn list_folders(
                 });
         }
     }
-    }
 
     folders.sort_by(|a, b| a.name.cmp(&b.name));
     Ok(Json(folders))

apps/api/src/libraries.rs

@@ -156,14 +156,19 @@ fn canonicalize_library_root(root_path: &str) -> Result<PathBuf, ApiError> {
         return Err(ApiError::bad_request("root_path must be absolute"));
     }
 
-    let canonical = std::fs::canonicalize(path)
+    // Avoid fs::canonicalize — it opens extra file descriptors to resolve symlinks
-        .map_err(|_| ApiError::bad_request("root_path does not exist or is inaccessible"))?;
+    // and can fail on Docker volume mounts (ro, cached) when fd limits are low.
-
+    if !path.exists() {
-    if !canonical.is_dir() {
+        return Err(ApiError::bad_request(format!(
+            "root_path does not exist: {}",
+            root_path
+        )));
+    }
+    if !path.is_dir() {
         return Err(ApiError::bad_request("root_path must point to a directory"));
     }
 
-    Ok(canonical)
+    Ok(path.to_path_buf())
 }
 
 use crate::index_jobs::{IndexJobResponse, RebuildRequest};

apps/api/src/pages.rs

@@ -16,7 +16,7 @@ use serde::Deserialize;
 use utoipa::ToSchema;
 use sha2::{Digest, Sha256};
 use sqlx::Row;
-use tracing::{debug, error, info, instrument, warn};
+use tracing::{error, info, instrument, warn};
 use uuid::Uuid;
 
 use crate::{error::ApiError, state::AppState};
@@ -32,9 +32,9 @@ fn remap_libraries_path(path: &str) -> String {
 
 fn parse_filter(s: &str) -> image::imageops::FilterType {
     match s {
-        "triangle" => image::imageops::FilterType::Triangle,
+        "lanczos3" => image::imageops::FilterType::Lanczos3,
         "nearest" => image::imageops::FilterType::Nearest,
-        _ => image::imageops::FilterType::Lanczos3,
+        _ => image::imageops::FilterType::Triangle, // Triangle (bilinear) is fast and good enough for comics
     }
 }
 
@@ -64,7 +64,7 @@ fn write_to_disk_cache(cache_path: &Path, data: &[u8]) -> Result<(), std::io::Er
     }
     let mut file = std::fs::File::create(cache_path)?;
     file.write_all(data)?;
-    file.sync_data()?;
+    // No sync_data() — this is a cache, durability is not critical
     Ok(())
 }
 
@@ -80,6 +80,8 @@ pub struct PageQuery {
 
 #[derive(Clone, Copy, Debug)]
 enum OutputFormat {
+    /// Serve raw bytes from the archive — no decode, no re-encode.
+    Original,
     Jpeg,
     Png,
     Webp,
@@ -87,16 +89,19 @@ enum OutputFormat {
 
 impl OutputFormat {
     fn parse(value: Option<&str>) -> Result<Self, ApiError> {
-        match value.unwrap_or("webp") {
+        match value {
-            "jpeg" | "jpg" => Ok(Self::Jpeg),
+            None => Ok(Self::Original),
-            "png" => Ok(Self::Png),
+            Some("original") => Ok(Self::Original),
-            "webp" => Ok(Self::Webp),
+            Some("jpeg") | Some("jpg") => Ok(Self::Jpeg),
-            _ => Err(ApiError::bad_request("format must be webp|jpeg|png")),
+            Some("png") => Ok(Self::Png),
+            Some("webp") => Ok(Self::Webp),
+            _ => Err(ApiError::bad_request("format must be original|webp|jpeg|png")),
         }
     }
 
     fn content_type(&self) -> &'static str {
         match self {
+            Self::Original => "application/octet-stream", // will be overridden by detected type
             Self::Jpeg => "image/jpeg",
             Self::Png => "image/png",
             Self::Webp => "image/webp",
@@ -105,6 +110,7 @@ impl OutputFormat {
 
     fn extension(&self) -> &'static str {
         match self {
+            Self::Original => "orig",
             Self::Jpeg => "jpg",
             Self::Png => "png",
             Self::Webp => "webp",
@@ -112,6 +118,17 @@ impl OutputFormat {
     }
 }
 
+/// Detect content type from raw image bytes.
+fn detect_content_type(data: &[u8]) -> &'static str {
+    match image::guess_format(data) {
+        Ok(ImageFormat::Jpeg) => "image/jpeg",
+        Ok(ImageFormat::Png) => "image/png",
+        Ok(ImageFormat::WebP) => "image/webp",
+        Ok(ImageFormat::Avif) => "image/avif",
+        _ => "application/octet-stream",
+    }
+}
+
 /// Get a specific page image from a book with optional format conversion
 #[utoipa::path(
     get,
@@ -132,30 +149,26 @@ impl OutputFormat {
     ),
     security(("Bearer" = []))
 )]
-#[instrument(skip(state), fields(book_id = %book_id, page = n))]
+#[instrument(skip(state, headers), fields(book_id = %book_id, page = n))]
 pub async fn get_page(
     State(state): State<AppState>,
     AxumPath((book_id, n)): AxumPath<(Uuid, u32)>,
     Query(query): Query<PageQuery>,
+    headers: HeaderMap,
 ) -> Result<Response, ApiError> {
-    info!("Processing image request");
-    
     if n == 0 {
-        warn!("Invalid page number: 0");
         return Err(ApiError::bad_request("page index starts at 1"));
     }
 
-    let (default_format, default_quality, max_width, filter_str, timeout_secs, cache_dir) = {
+    let (default_quality, max_width, filter_str, timeout_secs, cache_dir) = {
         let s = state.settings.read().await;
-        (s.image_format.clone(), s.image_quality, s.image_max_width, s.image_filter.clone(), s.timeout_seconds, s.cache_directory.clone())
+        (s.image_quality, s.image_max_width, s.image_filter.clone(), s.timeout_seconds, s.cache_directory.clone())
     };
 
-    let format_str = query.format.as_deref().unwrap_or(default_format.as_str());
+    let format = OutputFormat::parse(query.format.as_deref())?;
-    let format = OutputFormat::parse(Some(format_str))?;
     let quality = query.quality.unwrap_or(default_quality).clamp(1, 100);
     let width = query.width.unwrap_or(0);
     if width > max_width {
-        warn!("Invalid width: {}", width);
         return Err(ApiError::bad_request(format!("width must be <= {}", max_width)));
     }
     let filter = parse_filter(&filter_str);
@@ -165,11 +178,9 @@ pub async fn get_page(
 
     if let Some(cached) = state.page_cache.lock().await.get(&memory_cache_key).cloned() {
         state.metrics.page_cache_hits.fetch_add(1, Ordering::Relaxed);
-        debug!("Memory cache hit for key: {}", memory_cache_key);
+        return Ok(image_response(cached, format, None, &headers));
-        return Ok(image_response(cached, format.content_type(), None));
     }
     state.metrics.page_cache_misses.fetch_add(1, Ordering::Relaxed);
-    debug!("Memory cache miss for key: {}", memory_cache_key);
 
     let row = sqlx::query(
         r#"
@@ -191,7 +202,6 @@ pub async fn get_page(
     let row = match row {
         Some(r) => r,
         None => {
-            error!("Book file not found for book_id: {}", book_id);
             return Err(ApiError::not_found("book file not found"));
         }
     };
@@ -200,18 +210,22 @@ pub async fn get_page(
     let abs_path = remap_libraries_path(&abs_path);
     let input_format: String = row.get("format");
 
-    info!("Processing book file: {} (format: {})", abs_path, input_format);
-
     let disk_cache_key = get_cache_key(&abs_path, n, format.extension(), quality, width);
     let cache_path = get_cache_path(&disk_cache_key, &format, &cache_dir_path);
 
+    // If-None-Match: return 304 if the client already has this version
+    if let Some(if_none_match) = headers.get(header::IF_NONE_MATCH) {
+        let expected_etag = format!("\"{}\"", disk_cache_key);
+        if if_none_match.as_bytes() == expected_etag.as_bytes() {
+            return Ok(StatusCode::NOT_MODIFIED.into_response());
+        }
+    }
+
     if let Some(cached_bytes) = read_from_disk_cache(&cache_path) {
-        info!("Disk cache hit for: {}", cache_path.display());
         let bytes = Arc::new(cached_bytes);
         state.page_cache.lock().await.put(memory_cache_key, bytes.clone());
-        return Ok(image_response(bytes, format.content_type(), Some(&disk_cache_key)));
+        return Ok(image_response(bytes, format, Some(&disk_cache_key), &headers));
     }
-    debug!("Disk cache miss for: {}", cache_path.display());
 
     let _permit = state
         .page_render_limit
@@ -223,7 +237,6 @@ pub async fn get_page(
             ApiError::internal("render limiter unavailable")
         })?;
 
-    info!("Rendering page {} from {}", n, abs_path);
     let abs_path_clone = abs_path.clone();
     let format_clone = format;
     let start_time = std::time::Instant::now();
@@ -248,18 +261,27 @@ pub async fn get_page(
 
     match bytes {
         Ok(data) => {
-            info!("Successfully rendered page {} in {:?}", n, duration);
+            info!("Rendered page {} in {:?}", n, duration);
 
             if let Err(e) = write_to_disk_cache(&cache_path, &data) {
                 warn!("Failed to write to disk cache: {}", e);
-            } else {
-                info!("Cached rendered image to: {}", cache_path.display());
             }
 
             let bytes = Arc::new(data);
-            state.page_cache.lock().await.put(memory_cache_key, bytes.clone());
+            state.page_cache.lock().await.put(memory_cache_key.clone(), bytes.clone());
 
-            Ok(image_response(bytes, format.content_type(), Some(&disk_cache_key)))
+            // Prefetch next 2 pages in background (fire-and-forget)
+            for next_page in [n + 1, n + 2] {
+                let state2 = state.clone();
+                let abs_path2 = abs_path.clone();
+                let cache_dir2 = cache_dir_path.clone();
+                let format2 = format;
+                tokio::spawn(async move {
+                    prefetch_page(state2, book_id, &abs_path2, next_page, format2, quality, width, filter, timeout_secs, &cache_dir2).await;
+                });
+            }
+
+            Ok(image_response(bytes, format, Some(&disk_cache_key), &headers))
         }
         Err(e) => {
             error!("Failed to render page {} from {}: {:?}", n, abs_path, e);
@@ -268,11 +290,72 @@ pub async fn get_page(
     }
 }
 
-fn image_response(bytes: Arc<Vec<u8>>, content_type: &str, etag_suffix: Option<&str>) -> Response {
+/// Prefetch a single page into disk+memory cache (best-effort, ignores errors).
-    let mut headers = HeaderMap::new();
+async fn prefetch_page(
-    headers.insert(header::CONTENT_TYPE, HeaderValue::from_str(content_type).unwrap_or(HeaderValue::from_static("application/octet-stream")));
+    state: AppState,
-    headers.insert(header::CACHE_CONTROL, HeaderValue::from_static("public, max-age=31536000, immutable"));
+    book_id: Uuid,
+    abs_path: &str,
+    page: u32,
+    format: OutputFormat,
+    quality: u8,
+    width: u32,
+    filter: image::imageops::FilterType,
+    timeout_secs: u64,
+    cache_dir: &Path,
+) {
+    let mem_key = format!("{book_id}:{page}:{}:{quality}:{width}", format.extension());
+    // Already in memory cache?
+    if state.page_cache.lock().await.contains(&mem_key) {
+        return;
+    }
+    // Already on disk?
+    let disk_key = get_cache_key(abs_path, page, format.extension(), quality, width);
+    let cache_path = get_cache_path(&disk_key, &format, cache_dir);
+    if cache_path.exists() {
+        return;
+    }
+    // Acquire render permit (don't block too long — if busy, skip)
+    let permit = tokio::time::timeout(
+        Duration::from_millis(100),
+        state.page_render_limit.clone().acquire_owned(),
+    )
+    .await;
+    let _permit = match permit {
+        Ok(Ok(p)) => p,
+        _ => return,
+    };
 
+    // Fetch the book format from the path extension as a shortcut
+    let input_format = match abs_path.rsplit('.').next().map(|e| e.to_ascii_lowercase()) {
+        Some(ref e) if e == "cbz" => "cbz",
+        Some(ref e) if e == "cbr" => "cbr",
+        Some(ref e) if e == "pdf" => "pdf",
+        _ => return,
+    }
+    .to_string();
+
+    let abs_clone = abs_path.to_string();
+    let fmt = format;
+    let result = tokio::time::timeout(
+        Duration::from_secs(timeout_secs),
+        tokio::task::spawn_blocking(move || {
+            render_page(&abs_clone, &input_format, page, &fmt, quality, width, filter)
+        }),
+    )
+    .await;
+
+    if let Ok(Ok(Ok(data))) = result {
+        let _ = write_to_disk_cache(&cache_path, &data);
+        let bytes = Arc::new(data);
+        state.page_cache.lock().await.put(mem_key, bytes);
+    }
+}
+
+fn image_response(bytes: Arc<Vec<u8>>, format: OutputFormat, etag_suffix: Option<&str>, req_headers: &HeaderMap) -> Response {
+    let content_type = match format {
+        OutputFormat::Original => detect_content_type(&bytes),
+        _ => format.content_type(),
+    };
     let etag = if let Some(suffix) = etag_suffix {
         format!("\"{}\"", suffix)
     } else {
@@ -281,19 +364,37 @@ fn image_response(bytes: Arc<Vec<u8>>, content_type: &str, etag_suffix: Option<&
         format!("\"{:x}\"", hasher.finalize())
     };
 
+    // Check If-None-Match for 304
+    if let Some(if_none_match) = req_headers.get(header::IF_NONE_MATCH) {
+        if if_none_match.as_bytes() == etag.as_bytes() {
+            let mut headers = HeaderMap::new();
+            headers.insert(header::CACHE_CONTROL, HeaderValue::from_static("public, max-age=31536000, immutable"));
             if let Ok(v) = HeaderValue::from_str(&etag) {
                 headers.insert(header::ETAG, v);
             }
-    (StatusCode::OK, headers, Body::from((*bytes).clone())).into_response()
+            return (StatusCode::NOT_MODIFIED, headers).into_response();
+        }
+    }
+
+    let mut headers = HeaderMap::new();
+    headers.insert(header::CONTENT_TYPE, HeaderValue::from_str(content_type).unwrap_or(HeaderValue::from_static("application/octet-stream")));
+    headers.insert(header::CACHE_CONTROL, HeaderValue::from_static("public, max-age=31536000, immutable"));
+    if let Ok(v) = HeaderValue::from_str(&etag) {
+        headers.insert(header::ETAG, v);
+    }
+    // Use Bytes to avoid cloning the Vec — shares the Arc's allocation via zero-copy
+    let body_bytes = axum::body::Bytes::from(Arc::unwrap_or_clone(bytes));
+    (StatusCode::OK, headers, Body::from(body_bytes)).into_response()
 }
 
 /// Render page 1 of a book (for thumbnail fallback or thumbnail checkup). Uses thumbnail dimensions by default.
+/// Render page 1 as a thumbnail fallback. Returns (bytes, content_type).
 pub async fn render_book_page_1(
     state: &AppState,
     book_id: Uuid,
     width: u32,
     quality: u8,
-) -> Result<Vec<u8>, ApiError> {
+) -> Result<(Vec<u8>, &'static str), ApiError> {
     let row = sqlx::query(
         r#"SELECT abs_path, format FROM book_files WHERE book_id = $1 ORDER BY updated_at DESC LIMIT 1"#,
     )
@@ -328,7 +429,7 @@ pub async fn render_book_page_1(
                 &abs_path_clone,
                 &input_format,
                 1,
-                &OutputFormat::Webp,
+                &OutputFormat::Original,
                 quality,
                 width,
                 filter,
@@ -339,7 +440,9 @@ pub async fn render_book_page_1(
     .map_err(|_| ApiError::internal("page rendering timeout"))?
     .map_err(|e| ApiError::internal(format!("render task failed: {e}")))?;
 
-    bytes
+    let bytes = bytes?;
+    let content_type = detect_content_type(&bytes);
+    Ok((bytes, content_type))
 }
 
 fn render_page(
@@ -370,43 +473,93 @@ fn render_page(
         ApiError::internal(format!("page extraction failed: {e}"))
     })?;
 
+    // Original mode or source matches output with no resize → return raw bytes (zero transcoding)
+    if matches!(out_format, OutputFormat::Original) && width == 0 {
+        return Ok(page_bytes);
+    }
+    if width == 0 {
+        if let Ok(source_fmt) = image::guess_format(&page_bytes) {
+            if format_matches(&source_fmt, out_format) {
+                return Ok(page_bytes);
+            }
+        }
+    }
+
     transcode_image(&page_bytes, out_format, quality, width, filter)
 }
 
 
+/// Fast JPEG decode with DCT scaling: decodes directly at reduced resolution.
+fn fast_jpeg_decode(input: &[u8], target_w: u32, target_h: u32) -> Option<image::DynamicImage> {
+    if image::guess_format(input).ok()? != ImageFormat::Jpeg {
+        return None;
+    }
+    let mut decoder = jpeg_decoder::Decoder::new(std::io::Cursor::new(input));
+    decoder.read_info().ok()?;
+    decoder.scale(target_w as u16, target_h as u16).ok()?;
+    let pixels = decoder.decode().ok()?;
+    let info = decoder.info()?;
+    let w = info.width as u32;
+    let h = info.height as u32;
+    match info.pixel_format {
+        jpeg_decoder::PixelFormat::RGB24 => {
+            let buf = image::RgbImage::from_raw(w, h, pixels)?;
+            Some(image::DynamicImage::ImageRgb8(buf))
+        }
+        jpeg_decoder::PixelFormat::L8 => {
+            let buf = image::GrayImage::from_raw(w, h, pixels)?;
+            Some(image::DynamicImage::ImageLuma8(buf))
+        }
+        _ => None,
+    }
+}
+
 fn transcode_image(input: &[u8], out_format: &OutputFormat, quality: u8, width: u32, filter: image::imageops::FilterType) -> Result<Vec<u8>, ApiError> {
-    debug!("Transcoding image: {} bytes, format: {:?}, quality: {}, width: {}", input.len(), out_format, quality, width);
     let source_format = image::guess_format(input).ok();
-    debug!("Source format detected: {:?}", source_format);
+
-    let needs_transcode = source_format.map(|f| !format_matches(&f, out_format)).unwrap_or(true);
+    // Resolve "Original" to the actual source format for encoding
+    let effective_format = match out_format {
+        OutputFormat::Original => match source_format {
+            Some(ImageFormat::Png) => OutputFormat::Png,
+            Some(ImageFormat::WebP) => OutputFormat::Webp,
+            _ => OutputFormat::Jpeg, // default to JPEG for original resize
+        },
+        other => *other,
+    };
+
+    let needs_transcode = source_format.map(|f| !format_matches(&f, &effective_format)).unwrap_or(true);
 
     if width == 0 && !needs_transcode {
-        debug!("No transcoding needed, returning original");
         return Ok(input.to_vec());
     }
 
-    debug!("Loading image from memory...");
+    // For JPEG with resize: use DCT scaling to decode at ~target size (much faster)
-    let mut image = image::load_from_memory(input).map_err(|e| {
+    let mut image = if width > 0 {
-        error!("Failed to load image from memory: {} (input size: {} bytes)", e, input.len());
+        fast_jpeg_decode(input, width, u32::MAX)
+            .unwrap_or_else(|| {
+                image::load_from_memory(input).unwrap_or_default()
+            })
+    } else {
+        image::load_from_memory(input).map_err(|e| {
             ApiError::internal(format!("invalid source image: {e}"))
-    })?;
+        })?
+    };
 
     if width > 0 {
-        debug!("Resizing image to width: {}", width);
         image = image.resize(width, u32::MAX, filter);
     }
 
-    debug!("Converting to RGBA...");
     let rgba = image.to_rgba8();
     let (w, h) = rgba.dimensions();
-    debug!("Image dimensions: {}x{}", w, h);
 
     let mut out = Vec::new();
-    match out_format {
+    match effective_format {
-        OutputFormat::Jpeg => {
+        OutputFormat::Jpeg | OutputFormat::Original => {
+            // JPEG doesn't support alpha — convert RGBA to RGB
+            let rgb = image::DynamicImage::ImageRgba8(rgba.clone()).to_rgb8();
             let mut encoder = JpegEncoder::new_with_quality(&mut out, quality);
             encoder
-                .encode(&rgba, w, h, ColorType::Rgba8.into())
+                .encode(&rgb, w, h, ColorType::Rgb8.into())
                 .map_err(|e| ApiError::internal(format!("jpeg encode failed: {e}")))?;
         }
         OutputFormat::Png => {
@@ -421,7 +574,7 @@ fn transcode_image(input: &[u8], out_format: &OutputFormat, quality: u8, width:
                 .flat_map(|p| [p[0], p[1], p[2]])
                 .collect();
             let webp_data = webp::Encoder::new(&rgb_data, webp::PixelLayout::Rgb, w, h)
-                .encode(f32::max(quality as f32, 85.0));
+                .encode(quality as f32);
             out.extend_from_slice(&webp_data);
         }
     }

apps/api/src/state.rs

@@ -39,7 +39,7 @@ impl Default for DynamicSettings {
             timeout_seconds: 12,
             image_format: "webp".to_string(),
             image_quality: 85,
-            image_filter: "lanczos3".to_string(),
+            image_filter: "triangle".to_string(),
             image_max_width: 2160,
             cache_directory: std::env::var("IMAGE_CACHE_DIR")
                 .unwrap_or_else(|_| "/tmp/stripstream-image-cache".to_string()),

apps/backoffice/app/components/ui/Form.tsx

@@ -90,7 +90,7 @@ interface FormRowProps {
 }
 
 export function FormRow({ children, className = "" }: FormRowProps) {
-  return <div className={`flex flex-wrap items-end gap-4 ${className}`}>{children}</div>;
+  return <div className={`flex flex-wrap items-start gap-4 ${className}`}>{children}</div>;
 }
 
 // Form Section

apps/backoffice/app/settings/SettingsPage.tsx

@@ -88,13 +88,13 @@ export default function SettingsPage({ initialSettings, initialCacheStats, initi
             <Icon name="image" size="md" />
             Image Processing
           </CardTitle>
-          <CardDescription>Configure how images are processed and compressed</CardDescription>
+          <CardDescription>These settings only apply when a client explicitly requests format conversion via the API (e.g. <code className="text-xs bg-muted px-1 rounded">?format=webp&amp;width=800</code>). Pages served without parameters are delivered as-is from the archive, with no processing.</CardDescription>
         </CardHeader>
         <CardContent>
           <div className="space-y-4">
             <FormRow>
               <FormField className="flex-1">
-                <label className="text-sm font-medium text-muted-foreground mb-1 block">Output Format</label>
+                <label className="text-sm font-medium text-muted-foreground mb-1 block">Default Output Format</label>
                 <FormSelect
                   value={settings.image_processing.format}
                   onChange={(e) => {
@@ -103,13 +103,13 @@ export default function SettingsPage({ initialSettings, initialCacheStats, initi
                     handleUpdateSetting("image_processing", newSettings.image_processing);
                   }}
                 >
-                  <option value="webp">WebP (Recommended)</option>
+                  <option value="webp">WebP</option>
                   <option value="jpeg">JPEG</option>
                   <option value="png">PNG</option>
                 </FormSelect>
               </FormField>
               <FormField className="flex-1">
-                <label className="text-sm font-medium text-muted-foreground mb-1 block">Quality (1-100)</label>
+                <label className="text-sm font-medium text-muted-foreground mb-1 block">Default Quality (1-100)</label>
                 <FormInput 
                   type="number" 
                   min={1} 
@@ -126,7 +126,7 @@ export default function SettingsPage({ initialSettings, initialCacheStats, initi
             </FormRow>
             <FormRow>
               <FormField className="flex-1">
-                <label className="text-sm font-medium text-muted-foreground mb-1 block">Resize Filter</label>
+                <label className="text-sm font-medium text-muted-foreground mb-1 block">Default Resize Filter</label>
                 <FormSelect 
                   value={settings.image_processing.filter}
                   onChange={(e) => {
@@ -141,7 +141,7 @@ export default function SettingsPage({ initialSettings, initialCacheStats, initi
                 </FormSelect>
               </FormField>
               <FormField className="flex-1">
-                <label className="text-sm font-medium text-muted-foreground mb-1 block">Max Width (px)</label>
+                <label className="text-sm font-medium text-muted-foreground mb-1 block">Max Allowed Width (px)</label>
                 <FormInput 
                   type="number" 
                   min={100} 
@@ -344,10 +344,16 @@ export default function SettingsPage({ initialSettings, initialCacheStats, initi
                     handleUpdateSetting("thumbnail", newSettings.thumbnail);
                   }}
                 >
-                  <option value="webp">WebP (Recommended)</option>
+                  <option value="original">Original (No Re-encoding)</option>
+                  <option value="webp">WebP</option>
                   <option value="jpeg">JPEG</option>
                   <option value="png">PNG</option>
                 </FormSelect>
+                <p className="text-xs text-muted-foreground mt-1">
+                  {settings.thumbnail.format === "original"
+                    ? "Resizes to target dimensions, keeps source format (JPEG→JPEG). Much faster generation."
+                    : "Resizes and re-encodes to selected format."}
+                </p>
               </FormField>
             </FormRow>
             <FormRow>

apps/indexer/Cargo.toml

@@ -12,7 +12,7 @@ axum.workspace = true
 chrono.workspace = true
 futures = "0.3"
 image.workspace = true
-notify = "8"
+jpeg-decoder.workspace = true
 num_cpus.workspace = true
 parsers = { path = "../../crates/parsers" }
 rayon.workspace = true

apps/indexer/src/analyzer.rs

@@ -1,6 +1,6 @@
 use anyhow::Result;
 use futures::stream::{self, StreamExt};
-use image::GenericImageView;
+use image::{GenericImageView, ImageEncoder};
 use parsers::{analyze_book, BookFormat};
 use sqlx::Row;
 use std::path::Path;
@@ -14,6 +14,7 @@ use crate::{job::is_job_cancelled, utils, AppState};
 #[derive(Clone)]
 struct ThumbnailConfig {
     enabled: bool,
+    format: Option<String>,
     width: u32,
     height: u32,
     quality: u8,
@@ -24,6 +25,7 @@ struct ThumbnailConfig {
 async fn load_thumbnail_config(pool: &sqlx::PgPool) -> ThumbnailConfig {
     let fallback = ThumbnailConfig {
         enabled: true,
+        format: Some("webp".to_string()),
         width: 300,
         height: 400,
         quality: 80,
@@ -51,6 +53,11 @@ async fn load_thumbnail_config(pool: &sqlx::PgPool) -> ThumbnailConfig {
                     .get("enabled")
                     .and_then(|v| v.as_bool())
                     .unwrap_or(fallback.enabled),
+                format: value
+                    .get("format")
+                    .and_then(|v| v.as_str())
+                    .map(|s| s.to_string())
+                    .or_else(|| fallback.format.clone()),
                 width: value
                     .get("width")
                     .and_then(|v| v.as_u64())
@@ -100,22 +107,143 @@ async fn load_thumbnail_concurrency(pool: &sqlx::PgPool) -> usize {
     }
 }
 
+/// Detect the image format from raw bytes and return the corresponding file extension.
+fn detect_image_ext(data: &[u8]) -> &'static str {
+    match image::guess_format(data) {
+        Ok(image::ImageFormat::Png) => "png",
+        Ok(image::ImageFormat::WebP) => "webp",
+        _ => "jpg", // JPEG is the most common in comic archives
+    }
+}
+
+/// Fast JPEG decode with DCT scaling: decodes directly at reduced resolution (1/8, 1/4, 1/2).
+/// Returns (DynamicImage, original_width, original_height) or None if not JPEG / decode fails.
+fn fast_jpeg_decode(image_bytes: &[u8], target_w: u32, target_h: u32) -> Option<(image::DynamicImage, u32, u32)> {
+    // Only attempt for JPEG
+    if image::guess_format(image_bytes).ok()? != image::ImageFormat::Jpeg {
+        return None;
+    }
+
+    let mut decoder = jpeg_decoder::Decoder::new(std::io::Cursor::new(image_bytes));
+    // Read header to get original dimensions
+    decoder.read_info().ok()?;
+    let info = decoder.info()?;
+    let orig_w = info.width as u32;
+    let orig_h = info.height as u32;
+
+    // Request DCT-scaled decode (picks smallest scale >= requested size)
+    decoder.scale(target_w as u16, target_h as u16).ok()?;
+
+    let pixels = decoder.decode().ok()?;
+    let info = decoder.info()?;
+    let dec_w = info.width as u32;
+    let dec_h = info.height as u32;
+
+    let img = match info.pixel_format {
+        jpeg_decoder::PixelFormat::RGB24 => {
+            let buf = image::RgbImage::from_raw(dec_w, dec_h, pixels)?;
+            image::DynamicImage::ImageRgb8(buf)
+        }
+        jpeg_decoder::PixelFormat::L8 => {
+            let buf = image::GrayImage::from_raw(dec_w, dec_h, pixels)?;
+            image::DynamicImage::ImageLuma8(buf)
+        }
+        _ => return None,
+    };
+    Some((img, orig_w, orig_h))
+}
+
 fn generate_thumbnail(image_bytes: &[u8], config: &ThumbnailConfig) -> anyhow::Result<Vec<u8>> {
+    let t0 = std::time::Instant::now();
+
+    // Try fast JPEG DCT-scaled decode first (decodes directly at ~target size)
+    let (img, orig_w, orig_h) = if let Some(result) = fast_jpeg_decode(image_bytes, config.width, config.height) {
+        result
+    } else {
+        // Fallback for PNG/WebP/other formats
         let img = image::load_from_memory(image_bytes)
             .map_err(|e| anyhow::anyhow!("failed to load image: {}", e))?;
-    let (orig_w, orig_h) = img.dimensions();
+        let (ow, oh) = img.dimensions();
-    let ratio_w = config.width as f32 / orig_w as f32;
+        (img, ow, oh)
-    let ratio_h = config.height as f32 / orig_h as f32;
+    };
-    let ratio = ratio_w.min(ratio_h);
+    let t_decode = t0.elapsed();
-    let new_w = (orig_w as f32 * ratio) as u32;
+
-    let new_h = (orig_h as f32 * ratio) as u32;
+    // Don't upscale — clamp to original size
-    let resized = img.resize(new_w, new_h, image::imageops::FilterType::Triangle);
+    let target_w = config.width.min(orig_w);
+    let target_h = config.height.min(orig_h);
+
+    let t1 = std::time::Instant::now();
+    // thumbnail() is optimized for large downscale ratios (uses fast sampling)
+    let resized = img.thumbnail(target_w, target_h);
+    let (w, h) = resized.dimensions();
+    let t_resize = t1.elapsed();
+
+    let format = config.format.as_deref().unwrap_or("webp");
+    info!(
+        "[THUMBNAIL] {}x{} -> {}x{}  decode={:.0}ms  resize={:.0}ms  encode_format={}",
+        orig_w, orig_h, w, h,
+        t_decode.as_secs_f64() * 1000.0,
+        t_resize.as_secs_f64() * 1000.0,
+        format,
+    );
+
+    let t2 = std::time::Instant::now();
+    let result = match format {
+        "original" => {
+            // Re-encode in source format (fast JPEG encode instead of slow WebP)
+            let source_format = image::guess_format(image_bytes).unwrap_or(image::ImageFormat::Jpeg);
+            match source_format {
+                image::ImageFormat::Png => {
                     let rgba = resized.to_rgba8();
-    let (w, h) = rgba.dimensions();
+                    let mut buf = Vec::new();
-    let rgb_data: Vec<u8> = rgba.pixels().flat_map(|p| [p[0], p[1], p[2]]).collect();
+                    let encoder = image::codecs::png::PngEncoder::new(&mut buf);
+                    encoder.write_image(&rgba, w, h, image::ColorType::Rgba8.into())
+                        .map_err(|e| anyhow::anyhow!("png encode failed: {}", e))?;
+                    Ok(buf)
+                }
+                _ => {
+                    let rgb = resized.to_rgb8();
+                    let mut buf = Vec::new();
+                    let mut encoder = image::codecs::jpeg::JpegEncoder::new_with_quality(&mut buf, config.quality);
+                    encoder.encode(&rgb, w, h, image::ColorType::Rgb8.into())
+                        .map_err(|e| anyhow::anyhow!("jpeg encode failed: {}", e))?;
+                    Ok(buf)
+                }
+            }
+        }
+        "jpeg" | "jpg" => {
+            let rgb = resized.to_rgb8();
+            let mut buf = Vec::new();
+            let mut encoder = image::codecs::jpeg::JpegEncoder::new_with_quality(&mut buf, config.quality);
+            encoder.encode(&rgb, w, h, image::ColorType::Rgb8.into())
+                .map_err(|e| anyhow::anyhow!("jpeg encode failed: {}", e))?;
+            Ok(buf)
+        }
+        "png" => {
+            let rgba = resized.to_rgba8();
+            let mut buf = Vec::new();
+            let encoder = image::codecs::png::PngEncoder::new(&mut buf);
+            encoder.write_image(&rgba, w, h, image::ColorType::Rgba8.into())
+                .map_err(|e| anyhow::anyhow!("png encode failed: {}", e))?;
+            Ok(buf)
+        }
+        _ => {
+            // WebP (default)
+            let rgb = resized.to_rgb8();
+            let rgb_data: &[u8] = rgb.as_raw();
             let quality = config.quality as f32;
-    let webp_data = webp::Encoder::new(&rgb_data, webp::PixelLayout::Rgb, w, h).encode(quality);
+            let webp_data = webp::Encoder::new(rgb_data, webp::PixelLayout::Rgb, w, h).encode(quality);
             Ok(webp_data.to_vec())
+        }
+    };
+    let t_encode = t2.elapsed();
+    info!(
+        "[THUMBNAIL] encode={:.0}ms  total={:.0}ms  output_size={}KB",
+        t_encode.as_secs_f64() * 1000.0,
+        t0.elapsed().as_secs_f64() * 1000.0,
+        result.as_ref().map(|b| b.len() / 1024).unwrap_or(0),
+    );
+    result
 }
 
 /// Save raw image bytes (as extracted from the archive) without any processing.
@@ -127,23 +255,32 @@ fn save_raw_image(book_id: Uuid, raw_bytes: &[u8], directory: &str) -> anyhow::R
     Ok(path.to_string_lossy().to_string())
 }
 
-/// Resize the raw image and save it as a WebP thumbnail, overwriting the raw file.
+/// Resize the raw image and save it as a thumbnail, overwriting the raw file.
-fn resize_raw_to_webp(
+fn resize_raw_to_thumbnail(
     book_id: Uuid,
     raw_path: &str,
     config: &ThumbnailConfig,
 ) -> anyhow::Result<String> {
     let raw_bytes = std::fs::read(raw_path)
         .map_err(|e| anyhow::anyhow!("failed to read raw image {}: {}", raw_path, e))?;
-    let webp_bytes = generate_thumbnail(&raw_bytes, config)?;
+    info!("[THUMBNAIL] book={} raw_size={}KB", book_id, raw_bytes.len() / 1024);
+    let thumb_bytes = generate_thumbnail(&raw_bytes, config)?;
 
-    let webp_path = Path::new(&config.directory).join(format!("{}.webp", book_id));
+    let format = config.format.as_deref().unwrap_or("webp");
-    std::fs::write(&webp_path, &webp_bytes)?;
+    let ext = match format {
+        "original" => detect_image_ext(&raw_bytes),
+        "jpeg" | "jpg" => "jpg",
+        "png" => "png",
+        _ => "webp",
+    };
 
-    // Delete the raw file now that the WebP is written
+    let thumb_path = Path::new(&config.directory).join(format!("{}.{}", book_id, ext));
+    std::fs::write(&thumb_path, &thumb_bytes)?;
+
+    // Delete the raw file now that the thumbnail is written
     let _ = std::fs::remove_file(raw_path);
 
-    Ok(webp_path.to_string_lossy().to_string())
+    Ok(thumb_path.to_string_lossy().to_string())
 }
 
 fn book_format_from_str(s: &str) -> Option<BookFormat> {
@@ -465,7 +602,7 @@ pub async fn analyze_library_books(
 
                 let raw_path_clone = raw_path.clone();
                 let thumb_result = tokio::task::spawn_blocking(move || {
-                    resize_raw_to_webp(book_id, &raw_path_clone, &config)
+                    resize_raw_to_thumbnail(book_id, &raw_path_clone, &config)
                 })
                 .await;
 
@@ -554,18 +691,17 @@ pub async fn regenerate_thumbnails(
 
     let mut deleted_count = 0usize;
     for book_id in &book_ids_to_clear {
-        // Delete WebP thumbnail
+        // Delete thumbnail in any format (webp, jpg, png) + raw
-        let webp_path = Path::new(&config.directory).join(format!("{}.webp", book_id));
+        for ext in &["webp", "jpg", "png", "raw"] {
-        if webp_path.exists() {
+            let path = Path::new(&config.directory).join(format!("{}.{}", book_id, ext));
-            if let Err(e) = std::fs::remove_file(&webp_path) {
+            if path.exists() {
-                warn!("[ANALYZER] Failed to delete thumbnail {}: {}", webp_path.display(), e);
+                if let Err(e) = std::fs::remove_file(&path) {
-            } else {
+                    warn!("[ANALYZER] Failed to delete thumbnail {}: {}", path.display(), e);
+                } else if *ext != "raw" {
                     deleted_count += 1;
                 }
             }
-        // Delete raw file if it exists (interrupted previous run)
+        }
-        let raw_path = Path::new(&config.directory).join(format!("{}.raw", book_id));
-        let _ = std::fs::remove_file(&raw_path);
     }
     info!("[ANALYZER] Deleted {} thumbnail files for regeneration", deleted_count);
 
@@ -599,14 +735,10 @@ pub async fn cleanup_orphaned_thumbnails(state: &AppState) -> Result<()> {
         for entry in entries.flatten() {
             let file_name = entry.file_name();
             let file_name = file_name.to_string_lossy();
-            // Clean up both .webp and orphaned .raw files
+            // Clean up all thumbnail formats and orphaned .raw files
-            let stem = if let Some(s) = file_name.strip_suffix(".webp") {
+            let stem = [".webp", ".jpg", ".png", ".raw"]
-                Some(s.to_string())
+                .iter()
-            } else if let Some(s) = file_name.strip_suffix(".raw") {
+                .find_map(|ext| file_name.strip_suffix(ext).map(|s| s.to_string()));
-                Some(s.to_string())
-            } else {
-                None
-            };
             if let Some(book_id_str) = stem {
                 if let Ok(book_id) = Uuid::parse_str(&book_id_str) {
                     if !existing_book_ids.contains(&book_id) {

apps/indexer/src/job.rs

@@ -37,27 +37,57 @@ pub async fn cleanup_stale_jobs(pool: &PgPool) -> Result<()> {
     Ok(())
 }
 
+/// Job types that modify book/thumbnail data and must not run concurrently.
+const EXCLUSIVE_JOB_TYPES: &[&str] = &[
+    "rebuild",
+    "full_rebuild",
+    "scan",
+    "thumbnail_rebuild",
+    "thumbnail_regenerate",
+];
+
+/// Active statuses (job is still in progress, not just queued).
+const ACTIVE_STATUSES: &[&str] = &[
+    "running",
+    "extracting_pages",
+    "generating_thumbnails",
+];
+
 pub async fn claim_next_job(pool: &PgPool) -> Result<Option<(Uuid, Option<Uuid>)>> {
     let mut tx = pool.begin().await?;
 
+    // Check if any exclusive job is currently active
+    let has_active_exclusive: bool = sqlx::query_scalar(
+        r#"
+        SELECT EXISTS(
+            SELECT 1 FROM index_jobs
+            WHERE status = ANY($1)
+              AND type = ANY($2)
+        )
+        "#,
+    )
+    .bind(ACTIVE_STATUSES)
+    .bind(EXCLUSIVE_JOB_TYPES)
+    .fetch_one(&mut *tx)
+    .await?;
+
     let row = sqlx::query(
         r#"
         SELECT j.id, j.type, j.library_id
         FROM index_jobs j
         WHERE j.status = 'pending'
           AND (
-            (j.type IN ('rebuild', 'full_rebuild') AND NOT EXISTS (
+            -- Exclusive jobs: only if no other exclusive job is active
-                SELECT 1 FROM index_jobs
+            (j.type = ANY($1) AND NOT $2::bool)
-                WHERE status = 'running'
-                  AND type IN ('rebuild', 'full_rebuild')
-            ))
             OR
-            j.type NOT IN ('rebuild', 'full_rebuild')
+            -- Non-exclusive jobs (cbr_to_cbz): can always run
+            j.type != ALL($1)
           )
         ORDER BY
             CASE j.type
                 WHEN 'full_rebuild' THEN 1
                 WHEN 'rebuild' THEN 2
+                WHEN 'scan' THEN 2
                 ELSE 3
             END,
             j.created_at ASC
@@ -65,6 +95,8 @@ pub async fn claim_next_job(pool: &PgPool) -> Result<Option<(Uuid, Option<Uuid>)
         LIMIT 1
         "#,
     )
+    .bind(EXCLUSIVE_JOB_TYPES)
+    .bind(has_active_exclusive)
     .fetch_optional(&mut *tx)
     .await?;
 
@@ -74,30 +106,8 @@ pub async fn claim_next_job(pool: &PgPool) -> Result<Option<(Uuid, Option<Uuid>)
     };
 
     let id: Uuid = row.get("id");
-    let job_type: String = row.get("type");
     let library_id: Option<Uuid> = row.get("library_id");
 
-    if job_type == "rebuild" || job_type == "full_rebuild" {
-        let has_running_rebuild: bool = sqlx::query_scalar(
-            r#"
-            SELECT EXISTS(
-                SELECT 1 FROM index_jobs
-                WHERE status = 'running'
-                  AND type IN ('rebuild', 'full_rebuild')
-                  AND id != $1
-            )
-            "#,
-        )
-        .bind(id)
-        .fetch_one(&mut *tx)
-        .await?;
-
-        if has_running_rebuild {
-            tx.rollback().await?;
-            return Ok(None);
-        }
-    }
-
     sqlx::query(
         "UPDATE index_jobs SET status = 'running', started_at = NOW(), error_opt = NULL WHERE id = $1",
     )

apps/indexer/src/watcher.rs

@@ -1,34 +1,77 @@
 use anyhow::Result;
-use notify::{Event, RecommendedWatcher, RecursiveMode, Watcher};
 use sqlx::Row;
-use std::collections::HashMap;
+use std::collections::{HashMap, HashSet};
+use std::path::Path;
 use std::time::Duration;
-use tokio::sync::mpsc;
+use tracing::{error, info, trace, warn};
-use tracing::{error, info, trace};
 use uuid::Uuid;
+use walkdir::WalkDir;
 
 use crate::utils;
 use crate::AppState;
 
-pub async fn run_file_watcher(state: AppState) -> Result<()> {
+/// Check if any indexing job is currently active.
-    let (tx, mut rx) = mpsc::channel::<(Uuid, String)>(100);
+async fn has_active_jobs(pool: &sqlx::PgPool) -> bool {
+    sqlx::query_scalar::<_, bool>(
+        "SELECT EXISTS(SELECT 1 FROM index_jobs WHERE status IN ('pending', 'running', 'extracting_pages', 'generating_thumbnails'))",
+    )
+    .fetch_one(pool)
+    .await
+    .unwrap_or(false)
+}
 
-    // Start watcher refresh loop
+/// Snapshot: set of book file paths found in a library.
-    let refresh_interval = Duration::from_secs(30);
+type LibrarySnapshot = HashSet<String>;
+
+/// Walk a library directory and collect all book file paths.
+/// Walks sequentially to limit open file descriptors.
+fn snapshot_library(root_path: &str) -> LibrarySnapshot {
+    let mut files = HashSet::new();
+    let walker = WalkDir::new(root_path)
+        .follow_links(true)
+        .into_iter()
+        .filter_map(|e| e.ok());
+
+    for entry in walker {
+        if entry.file_type().is_file() && parsers::detect_format(entry.path()).is_some() {
+            files.insert(entry.path().to_string_lossy().to_string());
+        }
+    }
+    files
+}
+
+pub async fn run_file_watcher(state: AppState) -> Result<()> {
+    let poll_interval = Duration::from_secs(30);
     let pool = state.pool.clone();
 
-    tokio::spawn(async move {
+    // Stored snapshots per library — used to detect additions/removals between polls
+    let mut snapshots: HashMap<Uuid, LibrarySnapshot> = HashMap::new();
+    // Track which libraries we're watching (to detect config changes)
     let mut watched_libraries: HashMap<Uuid, String> = HashMap::new();
 
     loop {
-            // Get libraries with watcher enabled
+        tokio::time::sleep(poll_interval).await;
-            match sqlx::query(
+
-                "SELECT id, root_path FROM libraries WHERE watcher_enabled = TRUE AND enabled = TRUE"
+        // Skip if any job is active — avoid competing for file descriptors
+        if has_active_jobs(&pool).await {
+            trace!("[WATCHER] Skipping poll — job active");
+            continue;
+        }
+
+        // Fetch enabled libraries with watcher
+        let rows = match sqlx::query(
+            "SELECT id, root_path FROM libraries WHERE watcher_enabled = TRUE AND enabled = TRUE",
         )
         .fetch_all(&pool)
         .await
         {
-                Ok(rows) => {
+            Ok(rows) => rows,
+            Err(err) => {
+                error!("[WATCHER] Failed to fetch libraries: {}", err);
+                continue;
+            }
+        };
+
         let current_libraries: HashMap<Uuid, String> = rows
             .into_iter()
             .map(|row| {
@@ -39,109 +82,98 @@ pub async fn run_file_watcher(state: AppState) -> Result<()> {
             })
             .collect();
 
-                    // Check if we need to recreate watcher
+        // If library config changed, reset snapshots for removed/changed libraries
-                    let needs_restart = watched_libraries.len() != current_libraries.len()
+        if current_libraries != watched_libraries {
-                        || watched_libraries.iter().any(|(id, path)| {
+            let removed: Vec<Uuid> = watched_libraries
-                            current_libraries.get(id) != Some(path)
+                .keys()
-                        });
+                .filter(|id| !current_libraries.contains_key(id))
-
+                .copied()
-                    if needs_restart {
+                .collect();
-                        info!("[WATCHER] Restarting watcher for {} libraries", current_libraries.len());
+            for id in removed {
-
+                snapshots.remove(&id);
+            }
             if !current_libraries.is_empty() {
-                            let tx_clone = tx.clone();
+                info!(
-                            let libraries_clone = current_libraries.clone();
+                    "[WATCHER] Watching {} libraries (lightweight poll, {}s interval)",
-
+                    current_libraries.len(),
-                            match setup_watcher(libraries_clone, tx_clone) {
+                    poll_interval.as_secs()
-                                Ok(_new_watcher) => {
+                );
-                                    watched_libraries = current_libraries;
+            } else {
-                                    info!("[WATCHER] Watching {} libraries", watched_libraries.len());
+                info!("[WATCHER] No libraries to watch");
-                                }
-                                Err(err) => {
-                                    error!("[WATCHER] Failed to setup watcher: {}", err);
-                                }
-                            }
-                        }
-                    }
-                }
-                Err(err) => {
-                    error!("[WATCHER] Failed to fetch libraries: {}", err);
             }
+            watched_libraries = current_libraries.clone();
         }
 
-            tokio::time::sleep(refresh_interval).await;
+        // Poll each library sequentially to limit concurrent file descriptor usage
+        for (library_id, root_path) in &current_libraries {
+            if !Path::new(root_path).is_dir() {
+                warn!("[WATCHER] Library {} path not accessible: {}", library_id, root_path);
+                continue;
             }
-    });
 
-    // Process watcher events
+            // Re-check between libraries in case a job was created
-    while let Some((library_id, file_path)) = rx.recv().await {
+            if has_active_jobs(&pool).await {
-        info!("[WATCHER] File changed in library {}: {}", library_id, file_path);
+                trace!("[WATCHER] Job became active during poll, stopping");
+                break;
+            }
 
-        // Check if there's already a pending job for this library
+            let root_owned = root_path.clone();
-        match sqlx::query_scalar::<_, bool>(
+            let new_snapshot = tokio::task::spawn_blocking(move || snapshot_library(&root_owned))
-            "SELECT EXISTS(SELECT 1 FROM index_jobs WHERE library_id = $1 AND status IN ('pending', 'running'))"
+                .await
+                .unwrap_or_default();
+
+            let changed = match snapshots.get(library_id) {
+                Some(old_snapshot) => *old_snapshot != new_snapshot,
+                None => {
+                    // First scan — store baseline, don't trigger a job
+                    trace!(
+                        "[WATCHER] Initial snapshot for library {}: {} files",
+                        library_id,
+                        new_snapshot.len()
+                    );
+                    snapshots.insert(*library_id, new_snapshot);
+                    continue;
+                }
+            };
+
+            if changed {
+                info!(
+                    "[WATCHER] Changes detected in library {} ({})",
+                    library_id, root_path
+                );
+
+                // Check if a job already exists for this library
+                let job_exists = sqlx::query_scalar::<_, bool>(
+                    "SELECT EXISTS(SELECT 1 FROM index_jobs WHERE library_id = $1 AND status IN ('pending', 'running', 'extracting_pages', 'generating_thumbnails'))",
                 )
                 .bind(library_id)
-        .fetch_one(&state.pool)
+                .fetch_one(&pool)
                 .await
-        {
+                .unwrap_or(true);
-            Ok(exists) => {
+
-                if !exists {
+                if !job_exists {
-                    // Create a quick scan job
                     let job_id = Uuid::new_v4();
                     match sqlx::query(
-                        "INSERT INTO index_jobs (id, library_id, type, status) VALUES ($1, $2, 'rebuild', 'pending')"
+                        "INSERT INTO index_jobs (id, library_id, type, status) VALUES ($1, $2, 'rebuild', 'pending')",
                     )
                     .bind(job_id)
                     .bind(library_id)
-                    .execute(&state.pool)
+                    .execute(&pool)
                     .await
                     {
-                        Ok(_) => info!("[WATCHER] Created job {} for library {}", job_id, library_id),
+                        Ok(_) => info!(
+                            "[WATCHER] Created rebuild job {} for library {}",
+                            job_id, library_id
+                        ),
                         Err(err) => error!("[WATCHER] Failed to create job: {}", err),
                     }
                 } else {
-                    trace!("[WATCHER] Job already pending for library {}, skipping", library_id);
+                    trace!("[WATCHER] Job already active for library {}, skipping", library_id);
-                }
-            }
-            Err(err) => error!("[WATCHER] Failed to check existing jobs: {}", err),
                 }
             }
 
-    Ok(())
+            // Update snapshot
-}
+            snapshots.insert(*library_id, new_snapshot);
-
+        }
-fn setup_watcher(
+    }
-    libraries: HashMap<Uuid, String>,
-    tx: mpsc::Sender<(Uuid, String)>,
-) -> Result<RecommendedWatcher> {
-    let libraries_for_closure = libraries.clone();
-    
-    let mut watcher = notify::recommended_watcher(move |res: Result<Event, notify::Error>| {
-        match res {
-            Ok(event) => {
-                if event.kind.is_modify() || event.kind.is_create() || event.kind.is_remove() {
-                    for path in event.paths {
-                        if let Some((library_id, _)) = libraries_for_closure.iter().find(|(_, root)| {
-                            path.starts_with(root)
-                        }) {
-                            let path_str = path.to_string_lossy().to_string();
-                            if parsers::detect_format(&path).is_some() {
-                                let _ = tx.try_send((*library_id, path_str));
-                            }
-                        }
-                    }
-                }
-            }
-            Err(err) => error!("[WATCHER] Event error: {}", err),
-        }
-    })?;
-
-    // Actually watch the library directories
-    for root_path in libraries.values() {
-        info!("[WATCHER] Watching directory: {}", root_path);
-        watcher.watch(std::path::Path::new(root_path), RecursiveMode::Recursive)?;
-    }
-
-    Ok(watcher)
 }

crates/parsers/src/lib.rs

@@ -1,7 +1,8 @@
 use anyhow::{Context, Result};
+use std::collections::HashMap;
 use std::io::{Read, Write};
 use std::path::{Path, PathBuf};
-use std::sync::OnceLock;
+use std::sync::{Mutex, OnceLock};
 
 #[derive(Debug, Clone, Copy, PartialEq, Eq)]
 pub enum BookFormat {
@@ -527,13 +528,21 @@ pub fn extract_page(path: &Path, format: BookFormat, page_number: u32, pdf_rende
     }
 }
 
-fn extract_cbz_page(path: &Path, page_number: u32, allow_fallback: bool) -> Result<Vec<u8>> {
+/// Cache of sorted image names per archive path. Avoids re-listing and sorting on every page request.
-    let file = std::fs::File::open(path)
+static CBZ_INDEX_CACHE: OnceLock<Mutex<HashMap<PathBuf, Vec<String>>>> = OnceLock::new();
-        .with_context(|| format!("cannot open cbz: {}", path.display()))?;
-    let index = page_number as usize - 1;
 
-    match zip::ZipArchive::new(file) {
+fn cbz_index_cache() -> &'static Mutex<HashMap<PathBuf, Vec<String>>> {
-        Ok(mut archive) => {
+    CBZ_INDEX_CACHE.get_or_init(|| Mutex::new(HashMap::new()))
+}
+
+/// Get sorted image names from cache, or list + sort + cache them.
+fn get_cbz_image_index(path: &Path, archive: &mut zip::ZipArchive<std::fs::File>) -> Vec<String> {
+    {
+        let cache = cbz_index_cache().lock().unwrap();
+        if let Some(names) = cache.get(path) {
+            return names.clone();
+        }
+    }
     let mut image_names: Vec<String> = Vec::new();
     for i in 0..archive.len() {
         let entry = match archive.by_index(i) {
@@ -546,6 +555,21 @@ fn extract_cbz_page(path: &Path, page_number: u32, allow_fallback: bool) -> Resu
         }
     }
     image_names.sort_by(|a, b| natord::compare(a, b));
+    {
+        let mut cache = cbz_index_cache().lock().unwrap();
+        cache.insert(path.to_path_buf(), image_names.clone());
+    }
+    image_names
+}
+
+fn extract_cbz_page(path: &Path, page_number: u32, allow_fallback: bool) -> Result<Vec<u8>> {
+    let file = std::fs::File::open(path)
+        .with_context(|| format!("cannot open cbz: {}", path.display()))?;
+    let index = page_number as usize - 1;
+
+    match zip::ZipArchive::new(file) {
+        Ok(mut archive) => {
+            let image_names = get_cbz_image_index(path, &mut archive);
 
             let selected = image_names
                 .get(index)