fix(release): resolve style.css asar lookup with native separators

The shared-font check still handed asar a forward-slash path, which fails on Windows the same way the inventory listing did.
fix(release): unbreak packaging workflow on all three platforms
2026-11-10 16:46:06 -08:00 · 2026-09-22 02:29:28 -07:00 · 2026-09-22 02:15:49 -07:00 · 2026-09-22 01:29:58 -07:00 · 2026-09-21 20:53:26 -07:00 · 2026-09-21 20:40:57 -07:00
276 changed files with 11715 additions and 1586 deletions
@@ -13,6 +13,9 @@ on:
        required: true
      APPLE_TEAM_ID:
        required: true
+      # Project TMDB key baked into release artifacts; builds stay valid without it.
+      SUBMINER_TMDB_API_KEY:
+        required: false

 permissions:
  contents: read
@@ -69,6 +72,8 @@ jobs:

      - name: Build AppImage
        run: bun run build:appimage
+        env:
+          SUBMINER_TMDB_API_KEY: ${{ secrets.SUBMINER_TMDB_API_KEY }}

      - name: Build unversioned AppImage
        run: |
@@ -168,6 +173,7 @@ jobs:
      - name: Build signed + notarized macOS artifacts
        run: bun run build:mac
        env:
+          SUBMINER_TMDB_API_KEY: ${{ secrets.SUBMINER_TMDB_API_KEY }}
          CSC_LINK: ${{ secrets.CSC_LINK }}
          CSC_KEY_PASSWORD: ${{ secrets.CSC_KEY_PASSWORD }}
          APPLE_ID: ${{ secrets.APPLE_ID }}
@@ -248,6 +254,8 @@ jobs:

      - name: Build unsigned Windows artifacts
        run: bun run build:win:unsigned
+        env:
+          SUBMINER_TMDB_API_KEY: ${{ secrets.SUBMINER_TMDB_API_KEY }}

      - name: Smoke packaged runtime assets
        shell: bash
@@ -27,6 +27,7 @@ jobs:
      APPLE_ID: ${{ secrets.APPLE_ID }}
      APPLE_APP_SPECIFIC_PASSWORD: ${{ secrets.APPLE_APP_SPECIFIC_PASSWORD }}
      APPLE_TEAM_ID: ${{ secrets.APPLE_TEAM_ID }}
+      SUBMINER_TMDB_API_KEY: ${{ secrets.SUBMINER_TMDB_API_KEY }}

  release:
    needs: [package]
@@ -93,12 +93,20 @@ jobs:
          sudo ln -sf /usr/bin/lua5.4 /usr/local/bin/lua
          lua -v

-      - name: Test suite (source)
-        run: bun run test:fast
+      - name: Launcher unit and script suites
+        run: bun run test:launcher:unit:src && bun run test:scripts

      - name: Environment suite
        run: bun run test:env

+      - name: Upload launcher smoke artifacts (on failure)
+        if: failure()
+        uses: actions/upload-artifact@v4
+        with:
+          name: launcher-smoke
+          path: .tmp/launcher-smoke/**
+          if-no-files-found: ignore
+
      - name: Coverage suite (maintained source lane)
        run: bun run test:coverage:src

@@ -112,17 +120,6 @@ jobs:
      - name: Stats UI tests
        run: bun run test:stats

-      - name: Launcher smoke suite (source)
-        run: bun run test:launcher:smoke:src
-
-      - name: Upload launcher smoke artifacts (on failure)
-        if: failure()
-        uses: actions/upload-artifact@v4
-        with:
-          name: launcher-smoke
-          path: .tmp/launcher-smoke/**
-          if-no-files-found: ignore
-
      - name: Build (bundle)
        run: bun run build

@@ -28,6 +28,7 @@ jobs:
      APPLE_ID: ${{ secrets.APPLE_ID }}
      APPLE_APP_SPECIFIC_PASSWORD: ${{ secrets.APPLE_APP_SPECIFIC_PASSWORD }}
      APPLE_TEAM_ID: ${{ secrets.APPLE_TEAM_ID }}
+      SUBMINER_TMDB_API_KEY: ${{ secrets.SUBMINER_TMDB_API_KEY }}

  release:
    needs: [package]
@@ -22,7 +22,7 @@
        "@types/node": "^24.10.0",
        "@types/ws": "^8.18.1",
        "electron": "42.6.0",
-        "electron-builder": "26.15.3",
+        "electron-builder": "26.16.1",
        "esbuild": "^0.25.12",
        "eslint": "^10.8.0",
        "prettier": "^3.8.1",
@@ -36,9 +36,9 @@
  },
  "overrides": {
    "@xmldom/xmldom": "0.8.15",
-    "app-builder-lib": "26.15.3",
+    "app-builder-lib": "26.16.1",
    "brace-expansion": "5.0.9",
-    "electron-builder-squirrel-windows": "26.15.3",
+    "electron-builder-squirrel-windows": "26.16.1",
    "fast-uri": "3.1.6",
    "form-data": "4.0.6",
    "ip-address": "10.2.0",
@@ -181,7 +181,7 @@

    "@neon-rs/load": ["@neon-rs/load@0.0.4", "", {}, "sha512-kTPhdZyTQxB+2wpiRcFWrDcejc4JI6tkPuS7UZCG4l6Zvc5kU/gGQ/ozvHTh1XR5tS+UlfAfGuPajjzQjCiHCw=="],

-    "@noble/hashes": ["@noble/hashes@2.2.0", "", {}, "sha512-IYqDGiTXab6FniAgnSdZwgWbomxpy9FtYvLKs7wCUs2a8RkITG+DFGO1DM9cr+E3/RgADRpFjrKVaJ1z6sjtEg=="],
+    "@noble/hashes": ["@noble/hashes@1.8.0", "", {}, "sha512-jCs9ldd7NwzpgXDIf6P3+NrHh9/sD6CQdxHyjQI+h/6rDNo88ypBxxz45UDuZHz9r3tNz7N/VInSVoVdtXEI4A=="],

    "@peculiar/asn1-schema": ["@peculiar/asn1-schema@2.8.0", "", { "dependencies": { "@peculiar/utils": "^2.0.2", "asn1js": "^3.0.10", "tslib": "^2.8.1" } }, "sha512-7YT0U/ze0tF2QOBbE15gKZwy5tvgGyLRiRHLzhlbOpf7BT032oBSd0haZqXn5W6l26WLlu3dyxzjM+2638/z2Q=="],

@@ -243,7 +243,7 @@

    "ansi-styles": ["ansi-styles@4.3.0", "", { "dependencies": { "color-convert": "^2.0.1" } }, "sha512-zbB9rCJAT1rbjiVDb2hqKFHNYLxgtk8NURxZ3IZwD3F6NtxbXZQCnnSi1Lkx+IDohdPlFp222wVALIheZJQSEg=="],

-    "app-builder-lib": ["app-builder-lib@26.15.3", "", { "dependencies": { "@electron/asar": "3.4.1", "@electron/fuses": "^1.8.0", "@electron/get": "^3.0.0", "@electron/notarize": "2.5.0", "@electron/osx-sign": "1.3.3", "@electron/rebuild": "^4.0.4", "@electron/universal": "2.0.3", "@malept/flatpak-bundler": "^0.4.0", "@noble/hashes": "^2.2.0", "@peculiar/webcrypto": "^1.7.1", "@types/fs-extra": "9.0.13", "ajv": "^8.18.0", "asn1js": "^3.0.10", "async-exit-hook": "^2.0.1", "builder-util": "26.15.3", "builder-util-runtime": "9.7.0", "chromium-pickle-js": "^0.2.0", "ci-info": "4.3.1", "debug": "^4.3.4", "dotenv": "^16.4.5", "dotenv-expand": "^11.0.6", "ejs": "^3.1.8", "electron-publish": "26.15.3", "fs-extra": "^10.1.0", "hosted-git-info": "^4.1.0", "isbinaryfile": "^5.0.0", "jiti": "^2.4.2", "js-yaml": "^4.1.0", "json5": "^2.2.3", "lazy-val": "^1.0.5", "minimatch": "^10.2.5", "pkijs": "^3.4.0", "plist": "3.1.0", "proper-lockfile": "^4.1.2", "resedit": "^1.7.0", "semver": "~7.7.3", "tar": "^7.5.7", "temp-file": "^3.4.0", "tiny-async-pool": "1.3.0", "unzipper": "^0.12.3", "which": "^5.0.0" }, "peerDependencies": { "dmg-builder": "26.15.3", "electron-builder-squirrel-windows": "26.15.3" } }, "sha512-2VnyWkqsP5v5XbBhL3tD5Syx8iNPBYsoU7kY4S2fz7wg8Rj/nztWKCUzGKaFRTv0Xwf3/H058CR1Kvtd/3lRow=="],
+    "app-builder-lib": ["app-builder-lib@26.16.1", "", { "dependencies": { "@electron/asar": "3.4.1", "@electron/fuses": "^1.8.0", "@electron/get": "^3.0.0", "@electron/notarize": "2.5.0", "@electron/osx-sign": "1.3.3", "@electron/rebuild": "^4.0.4", "@electron/universal": "2.0.3", "@malept/flatpak-bundler": "^0.4.0", "@noble/hashes": "^1.8.0", "@peculiar/webcrypto": "^1.7.1", "@types/fs-extra": "9.0.13", "ajv": "^8.18.0", "asn1js": "^3.0.10", "async-exit-hook": "^2.0.1", "builder-util": "26.16.0", "builder-util-runtime": "9.7.0", "chromium-pickle-js": "^0.2.0", "ci-info": "4.3.1", "debug": "^4.3.4", "dotenv": "^16.4.5", "dotenv-expand": "^11.0.6", "ejs": "^3.1.8", "electron-publish": "26.16.0", "fs-extra": "^10.1.0", "hosted-git-info": "^4.1.0", "isbinaryfile": "^5.0.0", "jiti": "^2.4.2", "js-yaml": "^4.1.0", "json5": "^2.2.3", "lazy-val": "^1.0.5", "minimatch": "^10.2.5", "pkijs": "^3.4.0", "plist": "3.1.0", "proper-lockfile": "^4.1.2", "resedit": "^1.7.0", "semver": "~7.7.3", "tar": "^7.5.7", "temp-file": "^3.4.0", "tiny-async-pool": "1.3.0", "unzipper": "^0.12.3", "which": "^5.0.0" }, "peerDependencies": { "dmg-builder": "26.16.1", "electron-builder-squirrel-windows": "26.16.1" } }, "sha512-FhaO6YOup01ZfQW0Z6gt3AyukJjv1gW4uFK47jTgwcHZKqyN/fSlK2LqPf9tAeZYLP2bRJLDzeOkRImsw2X4Pg=="],

    "argparse": ["argparse@2.0.1", "", {}, "sha512-8+9WqebbFzpX9OR+Wa6O29asIogeRMzcGtAINdpMHHyAg10f05aSFVBbcEqGf/PXw1EjAZ+q2/bEBg3DvurK3Q=="],

@@ -273,7 +273,7 @@

    "buffer-from": ["buffer-from@1.1.2", "", {}, "sha512-E+XQCRwSbaaiChtv6k6Dwgc+bx+Bs6vuKJHHl5kox/BaKbhiXzqQOwK4cO22yElGp2OCmjwVhT3HmxgyPGnJfQ=="],

-    "builder-util": ["builder-util@26.15.3", "", { "dependencies": { "@types/debug": "^4.1.6", "builder-util-runtime": "9.7.0", "chalk": "^4.1.2", "cross-spawn": "^7.0.6", "debug": "^4.3.4", "fs-extra": "^10.1.0", "http-proxy-agent": "^7.0.0", "https-proxy-agent": "^7.0.0", "js-yaml": "^4.1.0", "sanitize-filename": "^1.6.3", "source-map-support": "^0.5.19", "stat-mode": "^1.0.0", "temp-file": "^3.4.0", "tiny-async-pool": "1.3.0" } }, "sha512-q2hn7Mbo2nFNkVekPiHFx6Nfo3hURmES3tfBn+k5Pqxl2RkmP3QGqZUhH/q9Pch/4G05NRhPjDlVj1O8q4Txvw=="],
+    "builder-util": ["builder-util@26.16.0", "", { "dependencies": { "@types/debug": "^4.1.6", "builder-util-runtime": "9.7.0", "chalk": "^4.1.2", "cross-spawn": "^7.0.6", "debug": "^4.3.4", "fs-extra": "^10.1.0", "http-proxy-agent": "^7.0.0", "https-proxy-agent": "^7.0.0", "js-yaml": "^4.1.0", "sanitize-filename": "^1.6.3", "source-map-support": "^0.5.19", "stat-mode": "^1.0.0", "temp-file": "^3.4.0", "tiny-async-pool": "1.3.0" } }, "sha512-RLyJhB7Si3YkzKR9ubQslWuXW3Vhs3CGe1i+SeixBZ0qTd1mk3XBmssvY22TlB6CS5blyko8Gu1JzpYk8UkYAg=="],

    "builder-util-runtime": ["builder-util-runtime@9.7.0", "", { "dependencies": { "debug": "^4.3.4", "sax": "^1.2.4" } }, "sha512-g/kR520giAFYkSXTzcmF3kqQq7wi8F6N6SzeDgZrqTBN+VHdmgWOyTdD1yD7AATDId/yXLvuP34CxW46/BwCdw=="],

@@ -335,7 +335,7 @@

    "discord-api-types": ["discord-api-types@0.38.49", "", {}, "sha512-XnqcWmnFZFAE8ZM8SHAw9DIV8D3Or00rMQ8iQLotrEA2PmXhl+ykaf6L6q4l474hrSUH1JaYcv+iOMRWp2p6Tg=="],

-    "dmg-builder": ["dmg-builder@26.15.3", "", { "dependencies": { "app-builder-lib": "26.15.3", "builder-util": "26.15.3", "fs-extra": "^10.1.0", "js-yaml": "^4.1.0" } }, "sha512-O3zJUFUYHJKgzPqioHxfxzBzlSC1eXCSr79gMSBKBP5AgjjpmrydMsMLotEg9fAJF36vdUncb+4ndRNxoPdlSQ=="],
+    "dmg-builder": ["dmg-builder@26.16.1", "", { "dependencies": { "app-builder-lib": "26.16.1", "builder-util": "26.16.0", "fs-extra": "^10.1.0", "js-yaml": "^4.1.0" } }, "sha512-pnI/3Qb24Uk+rMTgIUrsVUKosVgwmBUdF8Zeb8TexOSbpq8MWc7v6l+n+FrEqVkjNZwzBN+XpDS9ENgZ/rkWAw=="],

    "dotenv": ["dotenv@16.6.1", "", {}, "sha512-uBq4egWHTcTt33a72vpSG0z3HnPuIl6NqYcTrKEg2azoEyl2hpW0zqlxysq2pK9HlDIHyHyakeYaYnSAwd8bow=="],

@@ -349,11 +349,11 @@

    "electron": ["electron@42.6.0", "", { "dependencies": { "@electron-internal/extract-zip": "^1.0.1", "@electron/get": "^5.0.0", "@types/node": "^24.9.0" }, "bin": { "electron": "cli.js", "install-electron": "install.js" } }, "sha512-axGNgd+yCTg+vi1VEGrQqAj9WVWkePKwbICSAvMiT2eTaxhij9a/xhBHD6rXV8wrlW9ZfJzE5+xg752ImxrmTw=="],

-    "electron-builder": ["electron-builder@26.15.3", "", { "dependencies": { "app-builder-lib": "26.15.3", "builder-util": "26.15.3", "builder-util-runtime": "9.7.0", "chalk": "^4.1.2", "ci-info": "^4.2.0", "dmg-builder": "26.15.3", "fs-extra": "^10.1.0", "lazy-val": "^1.0.5", "simple-update-notifier": "2.0.0", "yargs": "^17.6.2" }, "bin": { "electron-builder": "./cli.js", "install-app-deps": "./install-app-deps.js" } }, "sha512-a1KM5heqS3gQCZzizXEI8RjJy3QVogULPdeSknt76uLDpBIW/HDGsMg/XgP0riP6PI9COsRvFITKKGDqA8fJxA=="],
+    "electron-builder": ["electron-builder@26.16.1", "", { "dependencies": { "app-builder-lib": "26.16.1", "builder-util": "26.16.0", "builder-util-runtime": "9.7.0", "chalk": "^4.1.2", "ci-info": "^4.2.0", "dmg-builder": "26.16.1", "fs-extra": "^10.1.0", "lazy-val": "^1.0.5", "simple-update-notifier": "2.0.0", "yargs": "^17.6.2" }, "bin": { "electron-builder": "./cli.js", "install-app-deps": "./install-app-deps.js" } }, "sha512-LrLK65QX5PUYYODXqp23FKrV7CILTtVY7mrJckNknO9jLNSMiqFkKbSMiDRw4CjOADMPVDdWLxY4mezOZWswxg=="],

-    "electron-builder-squirrel-windows": ["electron-builder-squirrel-windows@26.15.3", "", { "dependencies": { "app-builder-lib": "26.15.3", "builder-util": "26.15.3", "electron-winstaller": "5.4.0" } }, "sha512-Jc19XPV9y9+2bAdZPkXuVNGNIEFBq9poHC61l8Kv6FdK7DRG3+Ic0rerC0DXOaeHNz8yW0fg/JnF8GQROOF5MA=="],
+    "electron-builder-squirrel-windows": ["electron-builder-squirrel-windows@26.16.1", "", { "dependencies": { "app-builder-lib": "26.16.1", "builder-util": "26.16.0", "electron-winstaller": "5.4.0" } }, "sha512-w0y44wSaT1l6R7CAGmeHn4nHPfvzDyCAU1xJyi1w9SbPYJpYn76SmHDzqHf8Y7l91cPWTdPYBpGQtB2T5mJ08A=="],

-    "electron-publish": ["electron-publish@26.15.3", "", { "dependencies": { "@types/fs-extra": "^9.0.11", "aws4": "^1.13.2", "builder-util": "26.15.3", "builder-util-runtime": "9.7.0", "chalk": "^4.1.2", "form-data": "^4.0.5", "fs-extra": "^10.1.0", "lazy-val": "^1.0.5", "mime": "^2.5.2" } }, "sha512-g/2bn8YTavY4cuS5F+jOS7zmZbXXBV8KZ8yHKfJjFPoKtzBqrpCdNPxBd3tqdBwP7BVd0lGzf7Bk2s0KesWZ4Q=="],
+    "electron-publish": ["electron-publish@26.16.0", "", { "dependencies": { "@types/fs-extra": "^9.0.11", "aws4": "^1.13.2", "builder-util": "26.16.0", "builder-util-runtime": "9.7.0", "chalk": "^4.1.2", "form-data": "^4.0.5", "fs-extra": "^10.1.0", "lazy-val": "^1.0.5", "mime": "^2.5.2" } }, "sha512-Vt3KzQIiw9BImvNOYtndg9Mjki+tl4+1sQiC/+G5j8khWaENOJFWodiB+sUl6yyHwtd37avehskdtPw7f8y/+Q=="],

    "electron-updater": ["electron-updater@6.8.9", "", { "dependencies": { "builder-util-runtime": "9.7.0", "fs-extra": "^10.1.0", "js-yaml": "^4.1.0", "lazy-val": "^1.0.5", "lodash.escaperegexp": "^4.1.2", "lodash.isequal": "^4.5.0", "semver": "~7.7.3", "tiny-typed-emitter": "^2.1.0" } }, "sha512-ZhVxM9iGONUpZGI1FxdMRgJjUFXi7AYGVa5PwKlO1tV1/4zDxQmfKpXOHVztKrd6L9rLcFjERvi1Mf2vxyTkig=="],

@@ -1,5 +1,4 @@
 type: docs
 area: launcher

- Documented private-runtime launcher installation, simplified first-run setup, custom app locations, legacy launcher migration, and package-managed updates. Updated release-note download guidance for the bundled runtime and Windows launcher.
- Documented the bundled Bun runtime's MIT and LGPL licensing, where the notices live in the app, and the corresponding-source release asset. The AUR package now lists the bundled runtime licenses and installs their notices under `/usr/share/licenses/subminer-bin`, and `subminer-assets.tar.gz` includes the same notices.
+- Documented the bundled-runtime launcher install, legacy launcher migration, package-managed updates, and the bundled Bun runtime's MIT and LGPL notices (installed under `/usr/share/licenses/subminer-bin` by the AUR package and included in `subminer-assets.tar.gz`).
@@ -1,5 +1,4 @@
 type: changed
 area: launcher

- Made every installed and downloadable launcher use the Bun runtime included with SubMiner. Added a Windows `subminer.cmd` download, persistent private runtime caches for Linux and Windows, and migration for recognized legacy launchers. Deferred migrations survive overlapping update checks and remain pending when startup cannot safely refresh the launcher.
- Simplified first-run setup to a single optional launcher control, with runtime repair guidance shown only when needed.
+- Every installed and downloadable launcher now uses the Bun runtime bundled with SubMiner instead of a system Bun. Recognized legacy launchers migrate automatically, Windows gets a `subminer.cmd` download, and first-run setup is reduced to a single optional launcher control with runtime repair guidance shown only when needed.
@@ -0,0 +1,4 @@
+type: internal
+area: ci
+
+- Removed duplicate source and launcher smoke executions from the reusable quality gate while preserving every distinct test lane and failure artifact.
@@ -0,0 +1,4 @@
+type: internal
+area: verification
+
+- Replaced mislabeled dist source reruns with a small Electron-runtime smoke check for compiled stats startup, HTTP service, native SQLite, port conflicts, and cleanup.
@@ -1,6 +1,4 @@
 type: changed
 area: sync

- Sync uses compressed, incremental rsync transfers on compatible macOS and Linux machines, caching the last received snapshot per peer to reduce traffic on subsequent syncs. Cache helpers work through the launcher; older apps and launchers fall back to compressed transfers without an upload cache.
- Machines without compatible rsync, including Windows endpoints, automatically use compressed scp transfers.
- Rsync explicitly uses SSH and aborts transfers that exceed 30 minutes before merging.
+- Sync now uses compressed, incremental rsync transfers between compatible macOS and Linux machines, caching the last received snapshot per peer to cut traffic on later syncs. Machines without compatible rsync (including Windows) fall back to compressed scp, and older peers still work without the upload cache. Transfers abort after 30 minutes.
@@ -0,0 +1,4 @@
+type: fixed
+area: anki
+
+- Added `ankiConnect.fields.wordAudio` so word audio is read separately from the sentence-audio destination, fixing animated images that started moving immediately when `fields.audio` pointed to `SentenceAudio`.
@@ -0,0 +1,4 @@
+type: fixed
+area: config
+
+- Validate direct AnkiConnect, Kiku, and Senren settings before admitting them to runtime config, with warnings and defaults for invalid values.
@@ -0,0 +1,4 @@
+type: fixed
+area: jellyfin
+
+- Jellyfin streams no longer leak URL-derived titles or credential-bearing stream URLs into metadata lookups, Anki source fields, Discord presence, stats identities, or AniList retry keys. Previously cached credential-bearing parser metadata is cleaned up without touching watch history or library assignments.
@@ -0,0 +1,7 @@
+type: fixed
+area: jellyfin
+
+- Jellyfin playback, subtitle, artwork, and remote-control URLs now authenticate with the `ApiKey` query parameter instead of the legacy `X-Emby-*` headers, so the integration works on Jellyfin 12 where legacy authorization is disabled by default.
+- "Play on SubMiner" keeps working on Jellyfin 12: the cast-target websocket answers keep-alive requests and reconnects when the server stops replying instead of silently dying after about a minute.
+- The Jellyfin "now playing" bar clears when you close or finish a cast video instead of running on to the end of the episode.
+- Anki cards mined from Jellyfin playback get the episode title in the misc info field again instead of "Unknown media".
@@ -0,0 +1,4 @@
+type: fixed
+area: stats
+
+- Reject malformed resource IDs and partly invalid ID lists before stats library mutations or cover backfills run.
@@ -0,0 +1,4 @@
+type: fixed
+area: stats
+
+- Stats server port conflicts are reported through status notifications instead of crashing SubMiner. Startup and shutdown are also more robust: concurrent startup requests are shared, background stop no longer disconnects foreground dashboards, and shutdown bounds how long it waits for active requests.
@@ -0,0 +1,4 @@
+type: fixed
+area: anki
+
+- Closing the overlay while media timing review is still loading now cancels setup and modal retries, restores playback if the review paused it, and cleans up the hidden preview player.
@@ -0,0 +1,4 @@
+type: fixed
+area: anki
+
+- Treat `ankiConnect.media.maxMediaDuration: 0` as unlimited for stats dashboard mining, matching overlay mining and configuration.
@@ -1,4 +1,4 @@
 type: docs
 area: subtitles

- Explain how dialogue generation retains uncertain audible sections, why songs may also be transcribed, how detected speech starts guide long-passage cuts to reduce early subtitles, and why each passage uses a fresh Whisper process.
+- Documented subtitle generation: model choice, VAD behavior, reference-track guided splitting, fallback behavior, and known limits.
@@ -1,6 +1,7 @@
 type: added
 area: subtitles

- Generate local Japanese SRT subtitles with whisper.cpp from a standalone modal opened with Ctrl+Shift+G, the empty subtitle sidebar's generation button, or `subminer generate-subs`, with shared progress reporting, cancellation, safe output files, and automatic loading into the matching mpv video. The sidebar button hides while subtitle lines are loaded.
- Configure an existing multilingual model in Settings or choose an official multilingual model, including quantized variants, in the modal or launcher. The modal shows download sizes, speed and accuracy guidance, and a recommended starting model before explicitly downloading a verified SubMiner-managed model. Executable paths are optional overrides; empty fields find whisper-cli, ffmpeg, and ffprobe on PATH. The modal's Local tools check and the launcher name any missing executable and its setting before downloading a model or extracting audio, and generation confirms the destination directory grants write and search permissions up front.
- Optionally select Focus on spoken dialogue in the modal and use Download speech detection model to install the separate Silero model with progress and cancellation. The choice lasts for the session; a configured VAD model path sets the default. Retain uncertain audible sections so VAD rejection does not discard dialogue under music, accepting that songs may also be transcribed. Keep passages intact within Whisper's audio window, split longer passages near detected speech starts or quiet pauses with overlapping context to reduce early subtitle timing, and combine overlapping duplicate cues even when punctuation differs. Run each passage in a fresh Whisper process to prevent repeated-character output caused by state carried between files, at the cost of reloading the model per passage. Preserve original media timing and separate repeated dialogue.
+- Generate Japanese SRT subtitles locally with whisper.cpp from a modal (Ctrl+Shift+G), the empty subtitle sidebar's generation button, or `subminer generate-subs`, with progress, cancellation, and automatic loading into mpv.
+- Pick an official multilingual model (including quantized variants) with size and accuracy guidance and download it in-app, or point Settings at an existing model. `large-v3-turbo` is recommended when CUDA support is detected, `small` otherwise. whisper-cli, ffmpeg, and ffprobe are found on PATH unless overridden, and missing tools are named before any download starts.
+- Optional "Focus on spoken dialogue" mode uses a separately downloadable Silero VAD model, keeping uncertain audible sections so dialogue under music is not dropped (songs may be transcribed too).
+- Long passages are split near detected speech starts or quiet pauses, guided by an eligible embedded or external subtitle track already loaded in mpv when one is available, to reduce early subtitle timing. Each passage runs in a fresh Whisper process to avoid repeated-character output.
@@ -0,0 +1,4 @@
+type: docs
+area: jellyfin
+
+- Documented how Jellyfin media titles and stats identities keep stream credentials out of metadata.
@@ -0,0 +1,4 @@
+type: added
+area: jimaku
+
+- Jimaku modal gains Anime / Live action tabs so searches can pull from Jimaku's live action catalogue instead of only anime entries (Arrow Left / Right switch tabs).
@@ -0,0 +1,4 @@
+type: docs
+area: mining
+
+- Document independent screenshot selection in media timing review.
@@ -0,0 +1,4 @@
+type: added
+area: mining
+
+- Choose a still screenshot independently of the audio in media timing review, with a live preview, a separate time slider, and decoded-frame stepping. Local video and seekable remote streams such as Jellyfin are supported.
@@ -1,4 +1,4 @@
 type: added
 area: overlay

- The overlay discovers non-conflicting keyboard bindings from mpv defaults, input.conf, and loaded scripts in the background. SubMiner controls and explicitly disabled bindings take precedence. Discovered bindings stay session-only and do not appear in SubMiner's help menu.
+- The overlay picks up non-conflicting keyboard bindings from mpv defaults, `input.conf`, and loaded scripts. SubMiner controls and explicitly disabled bindings take precedence; discovered bindings are session-only and not listed in the help menu.
@@ -1,5 +1,4 @@
 type: changed
 area: release

- Reduced installer and unpacked app size by excluding documentation demo media, dependency source maps, TypeScript files, test and fixture directories, other development files, and unused Koffi platform binaries, and sharing the existing Japanese UI font across windows.
- Added package content checks, published size reports with release comparisons, and packaged asset/native-module smoke checks to the shared stable and prerelease build workflow. Size growth is reported without blocking releases.
+- Reduced installer and unpacked app size by dropping demo media, source maps, TypeScript sources, test fixtures, and unused Koffi binaries from the package, and sharing one Japanese UI font across windows. Release builds now publish package size reports with comparisons against the previous release.
@@ -0,0 +1,4 @@
+type: fixed
+area: overlay
+
+- Cancel pending Linux overlay window replacements during teardown so a delayed close callback cannot reopen the overlay.
@@ -0,0 +1,4 @@
+type: docs
+area: config
+
+- Clarified live-setting save feedback, mixed restart warnings, and subtitle-generation reload behavior.
@@ -0,0 +1,5 @@
+type: fixed
+area: config
+
+- Settings marked LIVE now use the same reload policy as save results, fixing false restart warnings for notifications and subtitle generation.
+- Mixed saves apply live changes and list only sections with changed fields that require a restart.
@@ -1,4 +1,4 @@
 type: added
 area: overlay

- Select dialogue across subtitle sidebar rows and copy it without timestamps using Ctrl/Cmd+C or the Copy button. Selection keeps the excerpt in view during playback and does not seek or require mining a card.
+- Select dialogue across subtitle sidebar rows and copy it without timestamps using Ctrl/Cmd+C or the Copy button. Selecting does not seek or require mining a card.
@@ -0,0 +1,4 @@
+type: changed
+area: stats
+
+- The stats server now rejects requests from non-loopback hosts and browser origins and requires `application/json` for mutation bodies. The in-app stats overlay loads from the local server so it shares the same protection. Reverse-proxied or Tailscale Serve dashboards are unsupported; scripts that POST must set a JSON content type.
@@ -0,0 +1,4 @@
+type: docs
+area: stats
+
+- Documented TMDB linking, provider reassignment, merge compatibility, and credential command caching.
@@ -0,0 +1,6 @@
+type: added
+area: stats
+
+- Live-action dramas and movies in the stats Library get posters, synopses, and titles from TMDB. Release builds include a project key; `tmdb.apiKey` (or `tmdb.apiKeyCommand`) overrides it and is required when running from source.
+- Titles AniList cannot match are looked up on TMDB automatically when the parsed filename matches a Japanese live-action title exactly; otherwise use the new **Link to TMDB** action. Entries linked to the same TMDB title merge into one card, and the Library kind selector gained a Live Action option.
+- Provider reassignment keeps the previous link and artwork if the replacement download fails. Merges and sync keep AniList and TMDB identities separate, and the merge dialog explains mixed selections instead of failing.
@@ -0,0 +1,4 @@
+type: docs
+area: anki
+
+- Documented the separate word-audio field mapping; existing animated images need regeneration to pick up the fix.
@@ -0,0 +1,4 @@
+type: changed
+area: yomitan
+
+- Updated bundled Yomitan with upstream 26.9.8 changes, including historical Japanese kana transformations, Ukrainian language support, and improvements to Anki duplicate searches and audio retrieval.
@@ -0,0 +1,4 @@
+type: docs
+area: stats
+
+- Document YouTube channel filtering and video statistics in the Library.
@@ -0,0 +1,5 @@
+type: added
+area: stats
+
+- YouTube channels are now a separate Library media kind. Existing channel entries migrate automatically without changing viewing history or manual video assignments.
+- Added All Titles, Anime, and YouTube Library filters. Channels stay out of AniList matching, season repair, and duplicate recommendations, and merges or video moves cannot mix an anime entry with a YouTube channel.
@@ -541,7 +541,7 @@
  // ==========================================
  // AnkiConnect Integration
  // Automatic Anki updates and media generation options.
-  // Hot-reload: ankiConnect.ai.enabled, media.normalizeAudio/mirrorMpvVolume/reviewTiming, knownWords, nPlusOne, fields.word/audio/image/sentence/miscInfo, behavior.autoUpdateNewCards, isLapis.sentenceCardModel, isKiku.fieldGrouping, isSenren.fieldGrouping, and lapisKiku.wordCardKind update live while SubMiner is running.
+  // Hot-reload: ankiConnect.ai.enabled, media.normalizeAudio/mirrorMpvVolume/reviewTiming, knownWords, nPlusOne, fields.word/audio/wordAudio/image/sentence/miscInfo, behavior.autoUpdateNewCards, isLapis.sentenceCardModel, isKiku.fieldGrouping, isSenren.fieldGrouping, and lapisKiku.wordCardKind update live while SubMiner is running.
  // Shared AI provider transport settings are read from top-level ai and typically require restart.
  // Most other AnkiConnect settings still require restart.
  // ==========================================
@@ -562,6 +562,7 @@
    "fields": {
      "word": "Expression", // Card field for the mined word or expression text.
      "audio": "ExpressionAudio", // Card field that receives generated sentence audio.
+      "wordAudio": "ExpressionAudio", // Existing word-audio field read to time the frozen first frame of animated images. This mapping is only used for synchronization.
      "image": "Picture", // Card field that receives the captured screenshot or animated image.
      "sentence": "Sentence", // Card field that receives the source sentence text.
      "miscInfo": "MiscInfo", // Card field that receives the miscellaneous info pattern (see ankiConnect.metadata.pattern).
@@ -590,7 +591,7 @@
      "reviewTiming": false, // Review and preview subtitle media timing before SubMiner creates or enriches a mined card. Values: true | false
      "audioPadding": 0, // Seconds of padding appended to both ends of generated sentence audio and animated AVIF clips.
      "fallbackDuration": 3, // Fallback clip duration in seconds when subtitle timing data is unavailable.
-      "maxMediaDuration": 30 // Maximum allowed media clip duration in seconds.
+      "maxMediaDuration": 30 // Maximum allowed media clip duration in seconds. 0 disables the cap.
    }, // Media setting.
    "knownWords": {
      "highlightEnabled": false, // Enable fast local highlighting for words already known in Anki. Values: true | false
@@ -658,6 +659,16 @@
    "maxSearchResults": 10 // Maximum TsukiHime search results returned.
  }, // TsukiHime subtitle search configuration for Japanese primary and configured secondary subtitles. No API key required.

+  // ==========================================
+  // TMDB
+  // TMDB (The Movie Database) metadata for live-action dramas and movies in the stats Library: posters, synopses, and grouping by show.
+  // Hot-reload: TMDB changes apply to the next TMDB request.
+  // ==========================================
+  "tmdb": {
+    "apiKey": "", // Your own TMDB API key or read access token for live-action posters and synopses in the stats Library. Release builds bundle a project key, so set this only to use your own quota or when running from source (free under Settings > API on themoviedb.org).
+    "apiKeyCommand": "" // Shell command that prints the TMDB API key to stdout. Used instead of apiKey to avoid storing the key in plain text.
+  }, // TMDB (The Movie Database) metadata for live-action dramas and movies in the stats Library: posters, synopses, and grouping by show.
+
  // ==========================================
  // YouTube Playback Settings
  // Defaults for managed subtitle language preferences and YouTube subtitle loading.
@@ -47,6 +47,8 @@ The update flow:

 Failed AniList updates are persisted to a retry queue on disk and retried with exponential backoff.

+Updates are skipped if the media path cannot produce a safe, nonempty identity. Invalid entries are discarded when loading or adding to the retry queue.
+
 | Parameter        | Value      |
 | ---------------- | ---------- |
 | Initial backoff  | 30 seconds |
@@ -69,6 +71,8 @@ SubMiner fetches cover art from AniList for display in the stats dashboard. When

 A no-match result is cached for 5 minutes before SubMiner retries, preventing repeated API calls for unrecognized media.

+When AniList has no match, SubMiner tries [TMDB](/configuration#tmdb) next so live-action dramas and movies get a poster and synopsis too. See [Immersion tracking](/immersion-tracking#library) for how live-action entries are grouped.
+
 If the automatic match is wrong, use **Change AniList Entry** on a title in the stats Library. Relinking rewrites the cached art for every episode of that title, and both the detail view and the Library grid pick up the new cover right away: the grid refetches after a relink, and cover responses carry an ETag and are revalidated on each request instead of being cached for a day.

 ## Rate limiting
@@ -19,6 +19,8 @@ This project is built primarily for [Kiku](https://kiku.youyoumu.my.id/) and [La

 AnkiConnect listens on `http://127.0.0.1:8765` by default. If you changed the port in AnkiConnect's settings, update `ankiConnect.url` in your SubMiner config.

+AnkiConnect and Kiku/Senren settings follow the [configuration validation rules](/configuration#configuration-file): invalid values produce a warning and fall back to the option's default. Use JSON booleans such as `true`, not strings such as `"true"`, and a positive number for `ankiConnect.pollingRate`.
+
 ## Auto-enrichment transport

 When you add a word via Yomitan, SubMiner detects the new card and fills in the sentence, audio, and image fields automatically. Two detection methods are available:
@@ -125,6 +127,7 @@ SubMiner maps its data to your Anki note fields. Configure these under `ankiConn
  "fields": {
    "word": "Expression",        // mined word / expression text
    "audio": "SentenceAudio",    // sentence audio clip cut from the video
+    "wordAudio": "ExpressionAudio", // existing Yomitan word audio, read for animation sync
    "image": "Picture",          // screenshot or animated clip
    "sentence": "Sentence",      // subtitle text
    "miscInfo": "MiscInfo"       // metadata (filename, timestamp)
@@ -136,6 +139,8 @@ SubMiner maps its data to your Anki note fields. Configure these under `ankiConn

 Field names are matched against your Anki note type case-insensitively (an exact match wins, then a lowercase comparison). If a configured field does not exist on the note type, SubMiner skips it without error.

+`fields.wordAudio` selects the existing dictionary-audio field used to calculate the animated image's opening freeze. This mapping only reads audio; `fields.audio` still controls where generated sentence audio is written. See [config.example.jsonc](/config.example.jsonc) for defaults.
+
 These mappings always control normal word-card enrichment, including Yomitan proxy/polling updates and manual clipboard updates. Enabling Lapis or Kiku does not replace the configured word-card sentence and audio fields with `Sentence` and `SentenceAudio`. The dedicated sentence-card and audio-card shortcuts still use those Lapis/Kiku field names.

 Two related options live alongside `fields`: `ankiConnect.deck` (target deck; empty falls back as described above) and `ankiConnect.tags` (tags added to mined cards, default `["SubMiner"]`; set `[]` to disable tagging). The `miscInfo` content is controlled by `ankiConnect.metadata.pattern` (default `[SubMiner] %f (%t)`; tokens: `%f` filename, `%F` filename with extension, `%t` timestamp, `%T` timestamp with milliseconds, `<br>` newline).
@@ -183,12 +188,18 @@ Output format: MP3 at 44100 Hz. If the video has multiple audio streams, SubMine

 The audio is uploaded to Anki's media folder and inserted as `[sound:audio_<timestamp>.mp3]`.

+Overlay and stats-dashboard mining use the same `media.maxMediaDuration` limit. See the [configuration example](/config.example.jsonc) for its default and how to disable the cap.
+
 Set `media.reviewTiming` to `true` to pause playback and check the clip before its media is generated. It applies to word, sentence, and audio cards.

 The review opens on the subtitle range plus your configured audio padding. Subtitles usually hang around after the dialogue has stopped, so once the waveform loads, an untouched clip end pulls back to just after the last speech in the line. The Line end rail still marks the original subtitle timing, Reset puts it back, and a line whose speech runs right through its end is left alone.

 **Adjusting the clip.** Drag either edge to trim, drag the middle to slide the whole clip without changing its length, or click anywhere on the waveform to snap the nearer edge there. A focused edge also moves with the arrow keys: 100 ms per press, or 500 ms with Shift. The 100 ms buttons do the same thing. Earlier and Later each reveal two more seconds of timeline without moving the selection.

+**Choosing the screenshot.** With still images enabled, drag the screenshot slider or use the Frame buttons to choose a video frame independently of the audio. Earlier and Later reveal more time for both sliders. The image follows the audio midpoint until you pick a frame, then stays fixed while you trim audio. Reset below the image restores automatic selection; the audio Reset affects only the audio. The screenshot slider also supports arrow keys, Home, and End.
+
+The picker supports local video and seekable remote streams, including Jellyfin, without seeking the main player. A selected frame must load before you can confirm; if it fails, choose another frame or Reset. The picker is hidden when image generation is disabled or animated AVIF is selected. Animated images continue to use the reviewed audio range.
+
 **Keys.** Space previews the selection with a playhead sweeping the clip. The preview ends when the hidden player has actually played the last sample, so Bluetooth output latency does not clip the tail. Enter confirms and Escape cancels.

 **The waveform.** SubMiner reads a center channel when one carries dialogue and falls back to a mono mix otherwise, keeps only the 250 to 3500 Hz speech band, and draws each slice's loudness against the clip's own noise floor. Steady background music flattens out and dialogue stands up, which makes it much easier to tell adjacent lines apart. The mined subtitle appears as a tinted band with labeled line-start and line-end rails. If waveform analysis fails, the timing controls still work.
@@ -203,6 +214,8 @@ Confirming writes the combined lines to the sentence field. Reset drops the adde

 Clipboard updates and stats-dashboard mining never open timing review. The option is off by default and hot-reloads. **Review Media Timing** in the runtime options palette (`Ctrl/Cmd+Shift+O`) toggles it for the current session.

+If SubMiner closes the overlay while a timing review is still loading, it cancels pending setup and modal retries and restores playback if the review paused it. A new timing review can start after the overlay reopens.
+
 ### Screenshots (static)

 A single frame is captured at the current playback position.
@@ -237,7 +250,7 @@ SubMiner can produce an animated AVIF spanning the subtitle duration instead of
 }
 ```

-Animated AVIF requires an AV1 encoder (`libaom-av1`, `libsvtav1`, or `librav1e`) in your FFmpeg build. Generation timeout is 60 seconds. `media.syncAnimatedImageToWordAudio` (default `true`) prepends a frozen first frame matching the existing word-audio duration, so the motion starts together with the sentence audio.
+Animated AVIF requires an AV1 encoder (`libaom-av1`, `libsvtav1`, or `librav1e`) in your FFmpeg build. Generation timeout is 60 seconds. `media.syncAnimatedImageToWordAudio` (default `true`) prepends a frozen first frame matching the existing audio duration in `fields.wordAudio`, so the motion starts together with the sentence audio. The freeze is baked into the image when mined; changing the mapping does not repair previously generated images.

 ### Behavior options

@@ -61,7 +61,7 @@ The Settings window preserves existing JSONC comments, trailing commas, and unre

 Secret fields do not display stored values. They show whether a value is configured; entering a new value writes it, and reset clears the explicit path. Prefer command-based secret options such as `jimaku.apiKeyCommand` when available.

-Saving validates the candidate config before writing. Live-reloadable changes are applied immediately; other changes return a restart-required banner in the window.
+Saving validates the candidate config before writing. Saving only fields marked **LIVE** shows "Saved. Live settings applied." If a save also changes fields that need a restart, the banner lists only the sections containing those changed fields. Live changes still apply in the same save.

 ## Configuration file

@@ -103,7 +103,7 @@ SubMiner watches the active config file (`config.jsonc` or `config.json`) while

 Hot-reloadable settings include subtitle appearance, sidebar controls, keybindings,
 shortcuts, notifications, logging level, selected source-language preferences,
-Jimaku/Subsync settings, AniSkip settings (`mpv.aniskipEnabled`, `mpv.aniskipButtonKey`),
+Jimaku/Subsync and subtitle-generation settings, AniSkip settings (`mpv.aniskipEnabled`, `mpv.aniskipButtonKey`),
 stats keys (`stats.toggleKey`, `stats.markWatchedKey`), the secondary-subtitle default
 mode, and the Anki deck, known-word, N+1, field, sentence-card, and Kiku options
 listed in the reference tables below.
@@ -156,6 +156,7 @@ The configuration file includes several main sections:

 - [**Jimaku**](#jimaku) - Jimaku API configuration and defaults
 - [**TsukiHime**](#tsukihime) - Multi-language subtitle search and download
+- [**TMDB**](#tmdb) - Posters and synopses for live-action dramas and movies in the stats Library
 - [**Subtitle Sync**](#subtitle-sync) - Sync current subtitle with `alass`/`ffsubsync`
 - [**AniList**](#anilist) - Optional post-watch progress updates
 - [**Yomitan**](#yomitan) - Reuse an external read-only Yomitan profile
@@ -976,6 +977,7 @@ This example is intentionally compact. The option table below documents availabl
 | `ankiConnect.deck`                                | string                                      | Restrict duplicate detection and card enrichment to this Anki deck. Leave empty to use the Yomitan mining deck when available. In Settings, this dropdown auto-fills and persists Yomitan's current mining deck when available. |
 | `fields.word`                                     | string                                      | Card field for mined word / expression text (default: `Expression`)                                                                                                                                                             |
 | `fields.audio`                                    | string                                      | Card field for the generated sentence audio clip (default: `ExpressionAudio`). Set this to a dedicated field such as `SentenceAudio` so it does not collide with the word audio Yomitan writes.                                  |
+| `fields.wordAudio`                                | string                                      | Existing word-audio field read for the animated image's opening freeze. Independent of the sentence-audio destination in `fields.audio`; this mapping does not write audio. See [config.example.jsonc](/config.example.jsonc) for defaults. |
 | `fields.image`                                    | string                                      | Card field for images (default: `Picture`)                                                                                                                                                                                      |
 | `fields.sentence`                                 | string                                      | Card field for sentences (default: `Sentence`)                                                                                                                                                                                  |
 | `fields.miscInfo`                                 | string                                      | Card field for metadata (default: `"MiscInfo"`, set to `null` to disable)                                                                                                                                                       |
@@ -996,7 +998,7 @@ This example is intentionally compact. The option table below documents availabl
 | `media.syncAnimatedImageToWordAudio`              | `true`, `false`                             | Whether animated AVIF includes an opening frame synced to sentence word-audio timing (default: `true`).                                                                                                                         |
 | `media.audioPadding`                              | number (seconds)                            | Optional padding around generated sentence media timing (default: `0`). Animated AVIF clips include the same padded source range as sentence audio.                                                                             |
 | `media.fallbackDuration`                          | number (seconds)                            | Default duration if timing unavailable (default: `3.0`)                                                                                                                                                                         |
-| `media.maxMediaDuration`                          | number (seconds)                            | Max duration for generated media from multi-line copy (default: `30`, `0` to disable)                                                                                                                                           |
+| `media.maxMediaDuration`                          | number (seconds)                            | Maximum generated clip duration for overlay and stats-dashboard mining. See the [configuration example](/config.example.jsonc) for the default and disabling the cap.                                                           |
 | `behavior.overwriteAudio`                         | `true`, `false`                             | Replace existing audio on updates; when `false`, new audio is appended/prepended using the configured media insert mode; manual clipboard updates always replace generated sentence audio (default: `true`)                     |
 | `behavior.overwriteImage`                         | `true`, `false`                             | Replace existing images on updates; when `false`, new images are appended/prepended using the configured media insert mode (default: `true`)                                                                                    |
 | `behavior.mediaInsertMode`                        | `"append"`, `"prepend"`                     | Where to insert new media when overwrite is off (default: `"append"`)                                                                                                                                                           |
@@ -1157,6 +1159,32 @@ The keyboard shortcut lives under `shortcuts.openTsukihime` (default `Ctrl+Shift

 See [TsukiHime Integration](/tsukihime-integration) for the modal workflow, language tabs, and troubleshooting.

+### TMDB
+
+TMDB (The Movie Database) supplies posters, synopses, and show grouping for live-action dramas and movies in the stats [Library](/immersion-tracking#library). AniList only covers anime, so TMDB is what gives live-action titles a cover and a description.
+
+Release builds ship with a project TMDB key, so nothing needs to be configured. Set your own key to use your own quota, or when running SubMiner from source, where no key is bundled. Create one for free under **Settings > API** on [themoviedb.org](https://www.themoviedb.org/settings/api); either the short API key or the long "API Read Access Token" works.
+
+```json
+{
+  "tmdb": {
+    "apiKey": "",
+    "apiKeyCommand": "cat ~/.tmdb_key"
+  }
+}
+```
+
+| Option               | Values | Description                                                                                        |
+| -------------------- | ------ | -------------------------------------------------------------------------------------------------- |
+| `tmdb.apiKey`        | string | Your own TMDB API key or read access token; overrides the bundled key (default: empty)             |
+| `tmdb.apiKeyCommand` | string | Shell command that prints the key to stdout, used instead of `apiKey` to keep it out of the config |
+
+Successful `apiKeyCommand` output is cached for the running client until `tmdb.apiKey` or `tmdb.apiKeyCommand` changes. Failed or empty command output uses the bundled key when available and waits 30 seconds before the next request can retry the command. Changing either credential setting resets this cooldown.
+
+Changes apply to the next TMDB request without a restart.
+
+This product uses the TMDB API but is not endorsed or certified by TMDB.
+
 ### Japanese subtitle generation

 Open the standalone modal with `Ctrl+Shift+G`, configurable through `shortcuts.openSubtitleGeneration`, or use the subtitle sidebar button. See [shortcuts](/shortcuts) for the shared mpv and overlay keybindings.
@@ -31,12 +31,42 @@ Episode completion for local `watched` state uses the shared `DEFAULT_MIN_WATCH_

 The same immersion data powers the stats dashboard.

+The browser dashboard and in-app stats overlay both load from the local HTTP server.
+The server accepts loopback hosts only and rejects requests from other browser origins,
+including opaque origins such as `file://`. API clients without a browser origin can
+still use the local API. Mutation requests with a body must use `application/json`;
+bodyless deletion and Anki browse requests remain supported. Requests rejected by the
+host or origin checks receive `403`; mutation bodies without a JSON content type
+receive `415`.
+
+Use the loopback dashboard URL directly. Reverse-proxied dashboards and Tailscale
+Serve URLs are unsupported because their host or browser origin is not the local
+server's origin. SSH stats synchronization is unchanged.
+
+Scripts sending a JSON body must include the content type. For example, this
+requests a duplicate-line cleanup preview without changing the database. Replace
+the port if you configured a different `stats.serverPort`:
+
+```bash
+curl http://127.0.0.1:6969/api/stats/maintenance/duplicate-lines \
+  -H 'Content-Type: application/json' \
+  -d '{"dryRun":true}'
+```
+
 - In-app overlay: focus the visible overlay, then press the key from `stats.toggleKey` (default: `` ` `` / `Backquote`).
 - Launcher command: run `subminer stats` to start the local stats server on demand (it also opens the dashboard in your browser when `stats.autoOpenBrowser` is enabled; the default is `false`).
 - Background server: run `subminer stats -b` to start or reuse a dedicated background stats daemon without keeping the launcher attached, and `subminer stats -s` to stop that daemon.
 - Maintenance commands: run `subminer stats cleanup` or `subminer stats cleanup -v` to backfill/repair vocabulary metadata (`headword`, `reading`, POS) and purge stale or excluded rows from `imm_words` on demand; `subminer stats cleanup -l` repairs lifetime summary tables non-destructively (recomputed from per-episode history, so lifetime totals older than the session retention window are kept); `subminer stats cleanup --duplicate-lines` collapses repeated lines left behind by typeset subtitles (see [Repeated Line Cleanup](#repeated-line-cleanup)). `subminer stats rebuild` and `subminer stats backfill` rebuild or backfill rollup data.
 - Browser page: open `http://127.0.0.1:6969` directly if the local stats server is already running.

+SubMiner waits for the local server to bind before reporting that the dashboard is available. If another process already uses the configured port, the command reports the startup error and the desktop app stays open. Opening the in-app dashboard also reports startup failures through your configured status notifications.
+
+`subminer stats -s` stops a background stats server or cancels a pending background start. It leaves a foreground-only server running, so an open in-app dashboard stays connected. Shutdown gives active HTTP requests one second to finish before closing their connections and finalizing stats.
+
+### Stats API resource IDs
+
+Resource IDs in URLs must be positive safe integers written as decimal digits without leading zeros, fractions, or exponent notation. ID lists in JSON bodies must contain positive safe integer numbers. Invalid IDs or list entries return `400` before any mutation; bulk requests do not apply just the valid subset. Pagination limits keep their existing rounding and bounds.
+
 ### Dashboard tabs

 #### Overview
@@ -53,14 +83,16 @@ Local files and Jellyfin items with detected season numbers are split into seaso

 When older stats already grouped multiple seasons under one series entry, SubMiner moves parsed episodes into the season-specific entries on startup and rebuilds the affected summaries.

+**Live-action dramas and movies.** Anime covers come from AniList, which has no live-action titles. A title that AniList cannot match is looked up on [TMDB](/configuration#tmdb) instead (release builds bundle a key; source builds need your own): only a Japanese-language, non-animated result whose known titles match the parsed filename exactly is accepted, and it supplies the poster, synopsis, English and Japanese titles, and episode count. If nothing matches automatically, open the title and use **Link to TMDB** to search and pick it by hand. A TMDB show spans all of its seasons, so entries that resolve to the same TMDB title are merged into one card regardless of the season folder they came from, and the merged season titles are remembered so later episodes land on the same card. The **All Titles** / **Anime** / **Live Action** / **YouTube** selector above the grid narrows the Library to one kind, and a title's detail view shows whether it is a drama or a movie. Linking a title to AniList again turns it back into an anime entry. Changing providers downloads the replacement cover before saving the new link; a failed download leaves the previous link and artwork intact. A title without a cover clears the previous artwork. Automatic TMDB matching leaves existing AniList links unchanged.
+
 Jellyfin stream URLs are normalized to stable item links before stats titles are shown, so playback query parameters are not displayed in the dashboard.

-When YouTube channel metadata is available, the Library tab groups videos by creator/channel and treats each tracked video as an episode-like entry inside that channel section.
+When YouTube channel metadata is available, the Library tab groups videos by creator/channel. Use the kind selector above the grid (**All Titles**, **Anime**, **Live Action**, **YouTube**) to filter the library. Channel pages show tracked videos and their stats without AniList controls. Existing channel entries are classified as YouTube automatically on startup, preserving viewing history and manual video assignments. Anime and YouTube entries with the same normalized title remain separate, including during stats sync. Channels are excluded from anime metadata matching, season repair, and duplicate recommendations.

 A library entry is identified by its parsed title plus any detected season, so the same show can end up on several cards when releases disagree about the title or omit the season tag. Two fixes are available:

- **Merge duplicates.** Hit **Select** above the grid, tick the cards that are the same show, and choose **Merge Selected**. Pick which entry to keep in the dialog; every episode moves onto it and the other cards are removed. Nothing is deleted, so sessions, mined cards and watch time all carry over. SubMiner remembers the merged title variants, so future episodes parsed with one of those names join the kept entry instead of recreating a duplicate card.
- **Move a single episode.** Hover an episode row in a title's episode list and use the **→** button to reassign it to another library entry. The correction is remembered, so later filename parsing or Jellyfin metadata cannot move that episode back. For local files, later episodes in the same directory inherit the correction when their detected seasons are compatible and every manual correction there points to the same entry; a file that parses to a title which already has its own library entry keeps that identity instead. Conflicting seasons or manual destinations are left for review. If the move empties the old entry, that card is removed and you are returned to the grid.
+- **Merge duplicates.** Hit **Select** above the grid, tick the cards that are the same show, and choose **Merge Selected**. Pick which entry to keep in the dialog; every episode moves onto it and the other cards are removed. Nothing is deleted, so sessions, mined cards and watch time all carry over. AniList-linked and TMDB-linked entries cannot be merged together, and YouTube channels cannot be merged with anime or live-action entries. SubMiner remembers the merged title variants, so future episodes parsed with one of those names join the kept entry instead of recreating a duplicate card.
+- **Move a single episode.** Hover an episode row in a title's episode list and use the **→** button to reassign it to another library entry. Anime and live-action entries are interchangeable here, but a YouTube video can only move between channels. The correction is remembered, so later filename parsing or Jellyfin metadata cannot move that episode back. For local files, later episodes in the same directory inherit the correction when their detected seasons are compatible and every manual correction there points to the same entry; a file that parses to a title which already has its own library entry keeps that identity instead. Conflicting seasons or manual destinations are left for review. If the move empties the old entry, that card is removed and you are returned to the grid.

 Once cover art resolves a series to an AniList entry, cards with compatible seasons are folded together automatically only when the searched title exactly matches an AniList title or synonym. A fuzzy result that points at an AniList entry already used by another card appears as a **Possible duplicate** review above the Library grid instead. Choose **Review merge** to compare the cards and pick which one to keep, or **Not duplicates** to dismiss that suggestion permanently. Entries with conflicting explicit season numbers are left alone rather than merged or suggested.

@@ -327,14 +359,14 @@ LIMIT ?;
 - Large-table reads are index-backed for `sample_ms`, session time windows, frequency-ranked words/kanji, and cover-art identity lookups.
 - Workload-dependent tuning knobs remain at defaults unless you change them: `cache_size`, `mmap_size`, `temp_store`, `auto_vacuum`.

-### Schema (v23)
+### Schema (v24)

 The exact schema version lives in `SCHEMA_VERSION` (`src/core/services/immersion-tracker/types.ts`) and is recorded in the `imm_schema_version` table.

 Core tables:

 - `imm_videos` - video key/title/source metadata
- `imm_anime` - anime/series metadata referenced by videos and lifetime tables
+- `imm_anime` - series or YouTube channel metadata referenced by videos and lifetime tables, including the media kind (`anime`, `live_action` or `youtube`) and the AniList or TMDB link
 - `imm_anime_title_aliases` - alternate titles that resolve to the same anime row
 - `imm_anime_merge_recommendations` - candidate duplicate-series merges surfaced in the dashboard
 - `imm_sessions` - session UUID, video reference, timing/status, final denormalized totals
@@ -12,7 +12,7 @@ This is the recommended way to use Jellyfin with SubMiner. A terminal-only optio

 ## Requirements

- A Jellyfin server plus your username and password
+- A Jellyfin server plus your username and password (Jellyfin 12, which disables legacy authorization by default, is supported)
 - SubMiner installed and running (see [Installation](/installation))
 - On Linux, the session token is stored with `gnome-libsecret` by default

@@ -52,10 +52,13 @@ From then on, pause / resume / seek / stop and audio or subtitle track changes y
 - **mpv launches automatically.** If mpv isn't already running when you cast, SubMiner starts it with SubMiner defaults and the bundled mpv plugin, so keybindings work right away.
 - **The overlay is managed by SubMiner,** so your configured `subtitleStyle` controls how subtitles look. Use the [overlay-toggle shortcut](/shortcuts) to hide it for a session.
 - **Resume works.** If Jellyfin has a saved position for the item, SubMiner seeks there on load.
+- **Titles and credentials stay separate.** AniList, character dictionaries, Anki source fields, and Discord presence use media titles, never authenticated stream URLs. If a usable title is unavailable, lookups are skipped and source fields show an unknown-media label. Stats identifies Jellyfin videos by server and item ID without the stream URL or API key.
 - **Direct play first.** When the source allows it and the container is in your direct-play allowlist, SubMiner streams the original file; otherwise it requests a transcoded stream from Jellyfin.
 - **Japanese subtitles are auto-selected,** preferring Jellyfin's default and embedded tracks over external sidecar files when several match.
 - **Downloaded subtitles keep their original timing.** SubMiner removes Jellyfin's server-selected subtitle stream from the mpv load URL, suppresses the mpv plugin's one-shot subtitle auto-selection and overlay auto-start for managed Jellyfin loads, stages the subtitle files exposed by Jellyfin without letting mpv auto-switch between tracks, resets mpv's subtitle delay to zero, then selects the Japanese track. SubMiner does not compare Japanese and English cue timelines or save an inferred delay.

+On startup, SubMiner clears cached anime parser metadata containing both API-key text and Jellyfin stream markers. Metadata containing only one of these is preserved.
+
 ## Settings

 All Jellyfin options live under **Settings → Integrations → Jellyfin** (open settings from the tray's **Open SubMiner Settings**). The ones that matter for casting:
@@ -1,6 +1,6 @@
 # Jimaku integration

-[Jimaku](https://jimaku.cc) is a community subtitle repository for anime, built from files other learners uploaded. SubMiner talks to the Jimaku API, so you search, browse, and download Japanese subtitle files from inside the overlay. No alt-tabbing, no moving files around. A downloaded track loads into mpv right away.
+[Jimaku](https://jimaku.cc) is a community subtitle repository for anime and Japanese live action, built from files other learners uploaded. SubMiner talks to the Jimaku API, so you search, browse, and download Japanese subtitle files from inside the overlay. No alt-tabbing, no moving files around. A downloaded track loads into mpv right away.

 ::: tip Prerequisite: a free API key
 You need a Jimaku account and an API key (a personal access string) before this feature works. Create an account at [jimaku.cc](https://jimaku.cc), copy your key, and add it to your config as shown under [Configuration](#configuration) below. Without a key, the search modal will report "Jimaku API key not set."
@@ -14,21 +14,23 @@ When you open the modal, SubMiner parses the current video filename to extract a

 From there:

-1. **Search** - SubMiner queries the Jimaku API with the parsed title. Results appear as a list of anime entries (Japanese and English names).
-2. **Browse entries** - Select an entry to load its available subtitle files, filtered by episode if one was detected.
-3. **Browse files** - Files show name, size, and last-modified date. If a language preference is configured, files are sorted accordingly (e.g., Japanese-tagged files first).
-4. **Download** - Selecting a file downloads it to the same directory as the video (or a temp directory for remote/streamed media) and loads it into mpv as a new subtitle track.
+1. **Pick a catalogue** - The **Anime** and **Live action** tabs at the top of the modal choose which Jimaku catalogue to search. Switching tabs re-runs the current search. The choice persists until SubMiner restarts.
+2. **Search** - SubMiner queries the Jimaku API with the parsed title. Results appear as a list of entries (Japanese and English names).
+3. **Browse entries** - Select an entry to load its available subtitle files, filtered by episode if one was detected.
+4. **Browse files** - Files show name, size, and last-modified date. If a language preference is configured, files are sorted accordingly (e.g., Japanese-tagged files first).
+5. **Download** - Selecting a file downloads it to the same directory as the video (or a temp directory for remote/streamed media) and loads it into mpv as a new subtitle track.

 If no files match the current episode filter, a "Show all files" button lets you broaden the search to all episodes for that entry.

 ### Modal keyboard shortcuts

-| Key | Action |
-| --- | --- |
-| `Enter` (in text field) | Search |
-| `Enter` (in list) | Select entry / download file |
-| `Arrow Up` / `Arrow Down` | Navigate entries or files |
-| `Escape` | Close modal |
+| Key                          | Action                                        |
+| ---------------------------- | --------------------------------------------- |
+| `Enter` (in text field)      | Search                                        |
+| `Enter` (in list)            | Select entry / download file                  |
+| `Arrow Up` / `Arrow Down`    | Navigate entries or files                     |
+| `Arrow Left` / `Arrow Right` | Switch between the Anime and Live action tabs |
+| `Escape`                     | Close modal                                   |

 ## Configuration

@@ -41,26 +43,26 @@ Add a `jimaku` section to your `config.jsonc`:
    "apiKeyCommand": "cat ~/.jimaku_key",
    "apiBaseUrl": "https://jimaku.cc",
    "languagePreference": "ja",
-    "maxEntryResults": 10
-  }
+    "maxEntryResults": 10,
+  },
 }
 ```

-| Option | Type | Default | Description |
-| --- | --- | --- | --- |
-| `jimaku.apiKey` | `string` | - | Jimaku API key (plaintext). Mutually exclusive with `apiKeyCommand`. |
-| `jimaku.apiKeyCommand` | `string` | - | Shell command that prints the API key to stdout. Useful for secret managers (e.g., `pass jimaku/api-key`). |
-| `jimaku.apiBaseUrl` | `string` | `"https://jimaku.cc"` | Base URL for the Jimaku API. Only change this if using a mirror or local instance. |
-| `jimaku.languagePreference` | `"ja"` \| `"en"` \| `"none"` | `"ja"` | Sort subtitle files by language tag. `"ja"` pushes Japanese-tagged files to the top; `"en"` does the same for English. `"none"` preserves the API order. |
-| `jimaku.maxEntryResults` | `number` | `10` | Maximum number of anime entries returned per search. |
+| Option                      | Type                         | Default               | Description                                                                                                                                              |
+| --------------------------- | ---------------------------- | --------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| `jimaku.apiKey`             | `string`                     | -                     | Jimaku API key (plaintext). Mutually exclusive with `apiKeyCommand`.                                                                                     |
+| `jimaku.apiKeyCommand`      | `string`                     | -                     | Shell command that prints the API key to stdout. Useful for secret managers (e.g., `pass jimaku/api-key`).                                               |
+| `jimaku.apiBaseUrl`         | `string`                     | `"https://jimaku.cc"` | Base URL for the Jimaku API. Only change this if using a mirror or local instance.                                                                       |
+| `jimaku.languagePreference` | `"ja"` \| `"en"` \| `"none"` | `"ja"`                | Sort subtitle files by language tag. `"ja"` pushes Japanese-tagged files to the top; `"en"` does the same for English. `"none"` preserves the API order. |
+| `jimaku.maxEntryResults`    | `number`                     | `10`                  | Maximum number of entries returned per search.                                                                                                           |

 The keyboard shortcut is configured separately under `shortcuts`:

 ```jsonc
 {
  "shortcuts": {
-    "openJimaku": "Ctrl+Shift+J"
-  }
+    "openJimaku": "Ctrl+Shift+J",
+  },
 }
 ```

@@ -541,7 +541,7 @@
  // ==========================================
  // AnkiConnect Integration
  // Automatic Anki updates and media generation options.
-  // Hot-reload: ankiConnect.ai.enabled, media.normalizeAudio/mirrorMpvVolume/reviewTiming, knownWords, nPlusOne, fields.word/audio/image/sentence/miscInfo, behavior.autoUpdateNewCards, isLapis.sentenceCardModel, isKiku.fieldGrouping, isSenren.fieldGrouping, and lapisKiku.wordCardKind update live while SubMiner is running.
+  // Hot-reload: ankiConnect.ai.enabled, media.normalizeAudio/mirrorMpvVolume/reviewTiming, knownWords, nPlusOne, fields.word/audio/wordAudio/image/sentence/miscInfo, behavior.autoUpdateNewCards, isLapis.sentenceCardModel, isKiku.fieldGrouping, isSenren.fieldGrouping, and lapisKiku.wordCardKind update live while SubMiner is running.
  // Shared AI provider transport settings are read from top-level ai and typically require restart.
  // Most other AnkiConnect settings still require restart.
  // ==========================================
@@ -562,6 +562,7 @@
    "fields": {
      "word": "Expression", // Card field for the mined word or expression text.
      "audio": "ExpressionAudio", // Card field that receives generated sentence audio.
+      "wordAudio": "ExpressionAudio", // Existing word-audio field read to time the frozen first frame of animated images. This mapping is only used for synchronization.
      "image": "Picture", // Card field that receives the captured screenshot or animated image.
      "sentence": "Sentence", // Card field that receives the source sentence text.
      "miscInfo": "MiscInfo", // Card field that receives the miscellaneous info pattern (see ankiConnect.metadata.pattern).
@@ -590,7 +591,7 @@
      "reviewTiming": false, // Review and preview subtitle media timing before SubMiner creates or enriches a mined card. Values: true | false
      "audioPadding": 0, // Seconds of padding appended to both ends of generated sentence audio and animated AVIF clips.
      "fallbackDuration": 3, // Fallback clip duration in seconds when subtitle timing data is unavailable.
-      "maxMediaDuration": 30 // Maximum allowed media clip duration in seconds.
+      "maxMediaDuration": 30 // Maximum allowed media clip duration in seconds. 0 disables the cap.
    }, // Media setting.
    "knownWords": {
      "highlightEnabled": false, // Enable fast local highlighting for words already known in Anki. Values: true | false
@@ -658,6 +659,16 @@
    "maxSearchResults": 10 // Maximum TsukiHime search results returned.
  }, // TsukiHime subtitle search configuration for Japanese primary and configured secondary subtitles. No API key required.

+  // ==========================================
+  // TMDB
+  // TMDB (The Movie Database) metadata for live-action dramas and movies in the stats Library: posters, synopses, and grouping by show.
+  // Hot-reload: TMDB changes apply to the next TMDB request.
+  // ==========================================
+  "tmdb": {
+    "apiKey": "", // Your own TMDB API key or read access token for live-action posters and synopses in the stats Library. Release builds bundle a project key, so set this only to use your own quota or when running from source (free under Settings > API on themoviedb.org).
+    "apiKeyCommand": "" // Shell command that prints the TMDB API key to stdout. Used instead of apiKey to avoid storing the key in plain text.
+  }, // TMDB (The Movie Database) metadata for live-action dramas and movies in the stats Library: posters, synopses, and grouping by show.
+
  // ==========================================
  // YouTube Playback Settings
  // Defaults for managed subtitle language preferences and YouTube subtitle loading.
@@ -29,11 +29,21 @@ With speech detection configured, SubMiner keeps detected speech and other audib

 Passages that fit within Whisper's 30-second audio window stay intact. Longer passages prefer nearby detected speech starts when choosing cuts, falling back to quiet pauses, with a small overlap to provide context. This reduces early subtitles caused by starting a clip well before its dialogue, while keeping all retained audio covered. Matching overlapping cues are combined even when punctuation differs; repeated dialogue at separate times remains separate. SubMiner runs each passage in a fresh Whisper process so decoder state from earlier audio cannot affect later passages. This reloads the model for each passage and can increase generation time. Subtitle cues stay within the supplied audio and retain each passage's position on the original timeline. Progress reports the current passage.

-This mode favors retaining dialogue over excluding music, so songs and background sounds may also produce subtitles. It can take longer than transcribing only VAD-approved speech. Whisper can still miss or misrecognize dialogue, and its timestamps remain estimates. Uncheck **Focus on spoken dialogue** to return to full-audio transcription for the session, or clear `vadModelPath` to change the default. A selected detector or model that fails stops generation with an error. Existing subtitles are preserved.
+This mode favors retaining dialogue over excluding music, so songs and background sounds may also produce subtitles. It can take longer than transcribing only VAD-approved speech. Whisper can still miss or misrecognize dialogue, and its timestamps remain estimates. Uncheck **Focus on spoken dialogue** to disable VAD for the session, or clear `vadModelPath` to change the default. Without a usable subtitle reference, disabling VAD returns to full-audio transcription. A selected detector or model that fails stops generation with an error. Existing subtitles are preserved.
+
+## Using loaded subtitles as timing references
+
+When generating for the video currently open in mpv, SubMiner automatically looks for a dialogue subtitle track among its embedded subtitles and loaded external SRT, ASS/SSA, or WebVTT files. It prefers English, then tracks labeled full or dialogue. Forced tracks, image subtitles, generated subtitles, and tracks whose titles or filenames identify signs, songs, lyrics, karaoke, or opening/ending subtitles are skipped. These checks rely on metadata; an unlabeled signs-only file cannot always be identified.
+
+The selected reference appears in generation progress. SubMiner reads its timestamps, including the active primary or secondary track's subtitle delay, and uses nearby cue starts to guide cuts in long audio passages. Short passages stay intact. The reference works with or without **Focus on spoken dialogue**. With that option enabled, reference starts take priority over VAD starts when choosing a nearby cut; VAD still helps identify speech. Audio outside reference cues remains eligible for transcription, and Whisper still supplies the Japanese text and final timestamps. The reference is assumed to be timed for the playing video; this does not automatically sync a mistimed reference.
+
+Unreadable or empty references are skipped in favor of another eligible loaded track. If none can be read, generation uses its normal audio timing. The launcher uses loaded references only when its input matches the video currently open in mpv; standalone generation keeps its existing behavior. It captures reference tracks and delays together with the initial audio selection, before checking or downloading a model. When relying on mpv's selected audio, it stops and asks you to retry if the media changes or cannot be verified during capture.

 ## Choosing a model

-Start with **small** for a balance of Japanese recognition quality and CPU time. This is a general starting recommendation, not a benchmark for your hardware. Tiny and base need less memory and usually finish sooner, with more recognition errors. Medium and large models favor accuracy but need more resources. Large-v3-turbo is optimized for speed compared with large-v3, with some accuracy tradeoff; actual performance depends on your CPU, GPU, whisper.cpp build, and audio.
+The modal recommends **large-v3-turbo** when it detects an NVIDIA GPU through `nvidia-smi` and the selected `whisper-cli` discovers an available CUDA device. Otherwise it recommends **small** for a balance of Japanese recognition quality and CPU time. The check works before downloading a model and falls back to small if a tool is missing, fails, times out, or reports an unrecognized result. Vulkan, AMD, and Apple GPU support do not qualify for the turbo recommendation. Checks are cached for up to 30 seconds; changing the Whisper executable path triggers a new check.
+
+The recommendation labels the model picker and explains the detected support. It does not change your configured model, current selection, external model path, or launcher's model choice. It also does not force a GPU backend during transcription. These are starting recommendations, not hardware benchmarks or guarantees that every model fits in available GPU memory. Tiny and base need less memory and usually finish sooner, with more recognition errors. Medium and large models favor accuracy but need more resources. Large-v3-turbo is optimized for speed compared with large-v3, with some accuracy tradeoff; actual performance depends on your CPU, GPU, whisper.cpp build, and audio.

 The picker includes whisper.cpp's official multilingual tiny, base, small, medium, large-v1, large-v2, large-v3, and large-v3-turbo downloads, including their available quantized variants. Quantized models use less disk space and memory, with possible accuracy loss. English-only `.en` models are excluded. See the [upstream model list](https://github.com/ggml-org/whisper.cpp/blob/master/models/download-ggml-model.sh) and [Whisper's model guidance](https://github.com/openai/whisper#available-models-and-languages).

@@ -47,7 +57,7 @@ A configured external Model Path takes precedence and hides the managed model pi
 4. Optionally check **Focus on spoken dialogue** and click **Download speech detection model** if prompted.
 5. Click **Generate subtitles**.

-The modal shows audio preparation, transcription, and saving progress. Percentages appear when the underlying tool reports them. **Cancel** stops the current operation. Closing the modal lets the job continue; reopening it shows the current progress or result.
+The modal adapts to the player window, using a wider layout when space allows and scrolling in smaller windows. It shows audio preparation, transcription, and saving progress. Percentages appear when the underlying tool reports them. **Cancel** stops the current operation. Closing the modal lets the job continue; reopening it shows the current progress or result.

 **Escape** or **Close** closes the modal using the same focus and overlay restoration as other SubMiner modals. Change or disable its shortcut with `shortcuts.openSubtitleGeneration` in Settings. Ctrl+G remains assigned to field grouping.

@@ -92,6 +92,10 @@ delay. It also works with no running
 SubMiner app or mpv instance when you provide a file path. Omit the path to use the current
 mpv file and selected audio track.

+When the input matches the current mpv file, generation automatically uses an eligible embedded
+or loaded external subtitle track as a timing reference, preferring English dialogue and skipping
+tracks marked as signs, songs, or forced. See [timing references](/subtitle-generation#using-loaded-subtitles-as-timing-references).
+
 ```bash
 subminer generate-subs video.mkv --download-model
 subminer generate-subs video.mkv --model-path ~/models/ggml-medium.bin
@@ -149,6 +149,10 @@ These settings come from `config.jsonc` (or built-in defaults); there are no CLI
 - **Secondary subtitle fails**: Secondary track failures never block playback. The primary subtitle loads independently.
 - **Native mpv secondary rendering**: Stays hidden during YouTube flows so the SubMiner overlay remains the visible secondary subtitle surface.

+## Viewing stats
+
+The stats Library groups tracked YouTube videos by channel when channel metadata is available. Select **YouTube** in the Library filter to see those channels separately from anime. Each channel page lists its videos, watch time, vocabulary, and mined cards. See [Immersion Tracking](/immersion-tracking#library).
+
 ## Related pages

 - [Usage - YouTube Playback](/usage#youtube-playback)
@@ -15,7 +15,10 @@

 Stable and prerelease workflows share `.github/workflows/package-release.yml`.
 Both callers explicitly pass the five required macOS signing/notarization
-secrets; `GITHUB_TOKEN` remains automatically available to the reusable workflow.
+secrets plus the optional `SUBMINER_TMDB_API_KEY` (the project TMDB key that
+`scripts/prepare-build-assets.mjs` stages into `dist/bundled-integration-keys.json`;
+artifacts built without it simply require users to set `tmdb.apiKey`).
+`GITHUB_TOKEN` remains automatically available to the reusable workflow.
 Each platform verifies its ASAR and external resources before signing, then
 measures the signed app and installers before upload. Missing runtime assets,
 foreign SQLite/Koffi binaries, duplicate UI fonts, demo media, source maps,
@@ -30,8 +30,11 @@ Update checks and startup launcher migration share a serialized update-state sto

 - `src/main/` owns composition, runtime setup, IPC wiring, and app lifecycle adapters.
 - `src/main/boot/` owns boot-phase assembly seams so `src/main.ts` can stay focused on lifecycle coordination and startup-path selection.
+- `src/main/runtime/linux-overlay-mode-runtime.ts` owns Linux fullscreen mode state and window replacement. App cleanup cancels pending replacements; `main.ts` supplies window creation and subtitle refresh hooks.
 - `src/core/services/` owns focused runtime services plus pure or side-effect-bounded logic.
 - `src/core/services/subtitle-generation*.ts` shares local whisper.cpp transcription, safe model downloads, and progress between the launcher and Electron. Optional dialogue mode retains both Silero-detected speech and other audible sections, omits confidently silent gaps, decodes passages independently, and restores original media timing. `src/main/runtime/subtitle-generation-runtime.ts` owns the overlay job lifecycle and only loads completed subtitles into the same local media; `src/shared/subtitle-generation*.ts` owns configuration, the multilingual model catalog, and IPC contracts. The overlay runtime retains a session model selection, validates picker requests through IPC, and keeps external model paths authoritative.
+- Subtitle model recommendations use bounded `nvidia-smi` and Whisper CUDA discovery probes in `subtitle-generation-acceleration.ts`. The overlay runtime caches results by executable path for 30 seconds and exposes acceleration status through the existing status IPC. Recommendations do not alter model selection or transcription arguments.
+- `subtitle-generation-reference.ts` ranks mpv's loaded text subtitle tracks, excludes signs/songs and forced references, and extracts timing hints with FFmpeg. The overlay and launcher snapshot references only for matching media, including active subtitle delays. Hints guide long-passage cuts with or without VAD; they never limit audio coverage or replace Whisper timestamps.
 - `src/renderer/` owns overlay rendering and input behavior.
 - `src/config/` owns config definitions, defaults, loading, and resolution.
 - `src/types/` owns shared cross-runtime contracts via domain entrypoints; `src/types.ts` stays a compatibility barrel.
@@ -30,6 +30,7 @@ Read when: you need to find the owner module for a behavior or test surface
 - Immersion sync: `src/core/services/stats-sync/`, bound by `src/main/sync-cli.ts`.
  `snapshot-transfer.ts` selects compressed rsync or scp. `transfer-cache.ts` atomically retains the last successfully received snapshot per hashed peer/database identity under the config directory's `sync-transfer-cache/`. Cache copies seed isolated transfer directories; rsync verifies reconstructed files before the existing merge engine runs. The `--make-temp` / `--remove-temp` helpers accept an internal `--transfer-cache` key, with a fallback for older peers that do not recognize it.
 - AniList tracking + character dictionary: `src/core/services/anilist/`, `src/main/runtime/composers/anilist-*`, `src/main/character-dictionary-runtime.ts`, `src/main/character-dictionary-runtime/`
+- TMDB live-action metadata: `src/core/services/tmdb/` (client + exact-title resolver), `src/core/services/immersion-tracker/live-action-link.ts` (links an entry to a TMDB title and merges other holders of the same title). The AniList cover-art fetcher calls the resolver as its fallback; `imm_anime.media_kind` marks the result and keeps the entry out of AniList season repair.
 - Jellyfin integration: `src/core/services/jellyfin*.ts`, `src/main/runtime/composers/jellyfin-*`
 - Window trackers: `src/window-trackers/`
 - Stats HTTP app: `src/core/services/stats-server.ts`, with route groups and shared route support
@@ -22,9 +22,15 @@ Read when: selecting the right verification lane for a change
  pull requests, stable tags, and prerelease tags. Keep common quality steps
  there instead of copying them into caller workflows.
 - The reusable gate installs Lua and runs `bun run test:env`, so the shipped mpv
-  plugin tests run for every pull request and tagged release.
+  plugin tests and launcher smoke run for every pull request and tagged release.
  Lua installation uses only the runner's Ubuntu package sources so unrelated
  third-party repository failures do not block the gate.
+- In the reusable gate, `test:coverage:src` is also the blocking execution of the
+  discovered `src/**` test lane. The coverage runner returns the failing test's
+  status, so CI does not rerun that lane through `test:fast`. Launcher unit and
+  script tests still run separately because they are outside the coverage lane.
+- Launcher smoke artifacts are uploaded after `test:env` fails. CI does not rerun
+  launcher smoke solely to collect the same artifacts.

 ## Default Handoff Gate

@@ -49,7 +55,7 @@ bun run docs:build
 - Internal KB, `AGENTS.md`, or `.agents/skills/**` changes: `bun run test:docs:kb`
 - Config/schema/defaults: `bun run test:config`, then `bun run generate:config-example` if template/defaults changed
 - Launcher/plugin: `bun run test:launcher` or `bun run test:env`
- Runtime-compat / compiled behavior: `bun run test:runtime:compat`
+- Runtime-compat / compiled behavior after `bun run build`: `bun run test:runtime:compat`
 - Stats dashboard UI: `bun run test:stats`
 - Build/release scripts (`scripts/**`): `bun run test:scripts`
 - Packaging: build the platform package, then run `bun run test:package <resources-directory>`.
@@ -62,11 +68,30 @@ bun run docs:build

 ## Coverage Reporting

- `bun run test:coverage:src` runs the maintained `test:src` lane through a sharded coverage runner: one Bun coverage process per test file, then merged LCOV output.
+- `bun run test:coverage:src` runs the same discovered `bun-src-full` membership as
+  `test:src` through a sharded coverage runner: one Bun coverage process per test
+  file, then merged LCOV output.
+- A failing coverage shard stops the runner with a nonzero status. Coverage is a
+  source test gate, not a report-only step.
 - Machine-readable output lands at `coverage/test-src/lcov.info`.
 - Every reusable quality-gate run uploads that LCOV file as the
  `coverage-test-src` artifact.

+## Compiled Runtime Smoke
+
+- `bun run test:smoke:dist` and its `test:runtime:compat` alias require an existing
+  full build and fail with the missing artifact paths when `dist/` or the stats UI
+  bundle is absent.
+- The check runs the emitted stats daemon under Electron's Node runtime. It opens
+  the production HTTP server, queries the overview endpoint through native
+  libsql-backed storage, and leaves an HTTP request body unfinished before
+  shutting the daemon down. It verifies a clean exit and that the port and
+  ownership state are released despite the unfinished request.
+- The check also occupies the configured port, requires startup to fail without
+  stale ownership state, releases the conflict, and verifies a clean retry.
+- This is not a full Electron UI startup check. It does not require a display and
+  makes no claims about renderer, window, tray, or mpv behavior.
+
 ## Dependency Audit Policy

 - `bun audit --audit-level high` blocks the reusable quality gate.
@@ -136,25 +136,28 @@ test('current mpv generation requires an identifiable selected audio track', asy
  }
 });

-test('explicit local file leaves Japanese track selection to the shared generator', async () => {
+test('explicit playing file leaves audio selection to the generator but inspects subtitle references', async () => {
  const f = fixture();
  await runGenerateSubtitlesCommand(f.context, f.deps);
  assert.equal(f.generations[0]?.audioStreamIndex, undefined);
  assert.equal(
    f.commands.some((command) => command[1] === 'track-list'),
-    false,
+    true,
  );
 });

 test('launcher does not load generated subtitles after mpv switches files', async () => {
  const f = fixture(['generate-subs']);
-  let pathRequests = 0;
+  let changed = false;
  f.deps.mpvCommand = async (_socket, command) => {
    f.commands.push(command);
-    if (command[1] === 'path')
-      return ++pathRequests === 1 ? '/media/episode.mkv' : '/media/next.mkv';
+    if (command[1] === 'path') return changed ? '/media/next.mkv' : '/media/episode.mkv';
    return [{ type: 'audio', selected: true, 'ff-index': 2 }];
  };
+  f.deps.generate = async () => {
+    changed = true;
+    return '/media/episode.ja.srt';
+  };
  await runGenerateSubtitlesCommand(f.context, f.deps);
  assert.equal(
    f.commands.some((command) => command[0] === 'sub-add'),
@@ -163,6 +166,86 @@ test('launcher does not load generated subtitles after mpv switches files', asyn
  assert.match(f.output.join(''), /Saved Japanese subtitles/);
 });

+test('current mpv generation rejects tracks when playback changes during capture', async () => {
+  for (const nextMedia of ['/media/next.mkv', null]) {
+    const f = fixture(['generate-subs']);
+    let media: string | null = '/media/episode.mkv';
+    let modelChecks = 0;
+    f.deps.mpvCommand = async (_socket, command) => {
+      if (command[1] === 'path') return media;
+      if (command[1] === 'track-list') {
+        media = nextMedia;
+        return [{ type: 'audio', selected: true, 'ff-index': 99 }];
+      }
+      return undefined;
+    };
+    f.deps.resolveModel = async () => {
+      modelChecks += 1;
+      return { kind: 'external', path: '/models/model.bin' };
+    };
+    await assert.rejects(
+      runGenerateSubtitlesCommand(f.context, f.deps),
+      /mpv media.*Run generate-subs again/,
+    );
+    assert.equal(f.generations.length, 0);
+    assert.equal(modelChecks, 0);
+  }
+});
+
+test('explicit media or audio stream remains usable when the mpv snapshot changes', async () => {
+  for (const options of [['/media/episode.mkv'], ['--audio-stream', '7']]) {
+    const f = fixture(['generate-subs', ...options]);
+    let media = '/media/episode.mkv';
+    f.deps.mpvCommand = async (_socket, command) => {
+      if (command[1] === 'path') return media;
+      if (command[1] === 'track-list') {
+        media = '/media/next.mkv';
+        return [
+          { type: 'audio', selected: true, 'ff-index': 99 },
+          { type: 'sub', lang: 'eng', 'ff-index': 100 },
+        ];
+      }
+      return undefined;
+    };
+    await runGenerateSubtitlesCommand(f.context, f.deps);
+    assert.equal(f.generations[0]?.mediaPath, '/media/episode.mkv');
+    assert.equal(
+      f.generations[0]?.audioStreamIndex,
+      options[0] === '--audio-stream' ? 7 : undefined,
+    );
+    assert.deepEqual(f.generations[0]?.references, []);
+  }
+});
+
+test('launcher captures loaded external references only for the matching media', async () => {
+  for (const matching of [true, false]) {
+    const f = fixture();
+    f.deps.mpvCommand = async (_socket, command) => {
+      if (command[1] === 'path') return matching ? '/media/episode.mkv' : '/media/other.mkv';
+      if (command[1] === 'working-directory') return '/mpv';
+      if (command[1] === 'track-list')
+        return [
+          { type: 'sub', external: true, 'external-filename': 'episode.en.signs.ass' },
+          { type: 'sub', external: true, 'external-filename': 'episode.en.srt' },
+        ];
+      return undefined;
+    };
+    await runGenerateSubtitlesCommand(f.context, f.deps);
+    assert.deepEqual(
+      f.generations[0]?.references,
+      matching
+        ? [
+            {
+              label: 'episode.en.srt',
+              delaySeconds: 0,
+              source: { kind: 'external', path: '/mpv/episode.en.srt' },
+            },
+          ]
+        : [],
+    );
+  }
+});
+
 test('explicit managed model overrides external config and downloads before generation', async () => {
  const f = fixture([
    'generate-subs',
@@ -191,6 +274,69 @@ test('explicit managed model overrides external config and downloads before gene
  assert.equal(f.generations.length, 1);
 });

+test('audio and subtitle timing use the initial mpv snapshot across model setup', async () => {
+  for (const changeDuring of ['resolve', 'download']) {
+    const f = fixture(['generate-subs', '--download-model']);
+    let changed = false;
+    let trackReads = 0;
+    f.deps.mpvCommand = async (_socket, command) => {
+      if (command[1] === 'path') return '/media/episode.mkv';
+      if (command[1] === 'track-list') {
+        trackReads += 1;
+        return [
+          { type: 'audio', selected: true, 'ff-index': changed ? 3 : 2 },
+          { type: 'sub', id: 1, title: 'English Full', lang: 'eng', 'ff-index': changed ? 5 : 4 },
+        ];
+      }
+      if (command[1] === 'sid') return 1;
+      if (command[1] === 'sub-delay') return changed ? 9 : 1.5;
+      return undefined;
+    };
+    f.deps.resolveModel = async () => {
+      if (changeDuring === 'resolve') changed = true;
+      return { kind: 'missing', path: '/models/model.bin' };
+    };
+    f.deps.downloadModel = async () => {
+      changed = true;
+      return '/models/model.bin';
+    };
+    await runGenerateSubtitlesCommand(f.context, f.deps);
+    assert.equal(f.generations[0]?.audioStreamIndex, 2);
+    assert.deepEqual(f.generations[0]?.references, [
+      {
+        label: 'English Full',
+        delaySeconds: 1.5,
+        source: { kind: 'embedded', streamIndex: 4 },
+      },
+    ]);
+    assert.equal(trackReads, 1);
+  }
+});
+
+test('explicit audio stream bypasses mpv audio selection while retaining subtitle references', async () => {
+  const f = fixture(['generate-subs', '--audio-stream', '7']);
+  f.deps.mpvCommand = async (_socket, command) => {
+    if (command[1] === 'path') return '/media/episode.mkv';
+    if (command[1] === 'track-list')
+      return [
+        { type: 'audio', selected: true, external: true, 'ff-index': 0 },
+        { type: 'sub', id: 2, title: 'English Full', lang: 'eng', 'ff-index': 4 },
+      ];
+    if (command[1] === 'secondary-sid') return 2;
+    if (command[1] === 'secondary-sub-delay') return -0.5;
+    return undefined;
+  };
+  await runGenerateSubtitlesCommand(f.context, f.deps);
+  assert.equal(f.generations[0]?.audioStreamIndex, 7);
+  assert.deepEqual(f.generations[0]?.references, [
+    {
+      label: 'English Full',
+      delaySeconds: -0.5,
+      source: { kind: 'embedded', streamIndex: 4 },
+    },
+  ]);
+});
+
 test('generation can run standalone and never loads subtitles into another video', async () => {
  for (const playing of [null, '/media/different.mkv']) {
    const f = fixture();
@@ -8,6 +8,10 @@ import {
  resolveSubtitleGenerationTools,
 } from '../../src/core/services/subtitle-generation.js';
 import { requireSubtitleGenerationTools } from '../../src/core/services/subtitle-generation-tools.js';
+import {
+  readSubtitleGenerationReferences,
+  type SubtitleGenerationReference,
+} from '../../src/core/services/subtitle-generation-reference.js';
 import {
  resolveSubtitleGenerationConfig,
  type SubtitleGenerationProgress,
@@ -69,12 +73,9 @@ async function readMpvMedia(socketPath: string, command: GenerationCommandDeps['
  return localMediaPath(media, workingDirectory);
 }

-async function readMpvAudioStream(
-  socketPath: string,
-  command: GenerationCommandDeps['mpvCommand'],
-) {
-  const tracks = await command(socketPath, ['get_property', 'track-list'], 1000);
-  for (const track of Array.isArray(tracks) ? tracks : []) {
+function selectedMpvAudioStream(value: unknown) {
+  const tracks: unknown[] = Array.isArray(value) ? value : [];
+  for (const track of tracks) {
    if (
      typeof track === 'object' &&
      track !== null &&
@@ -160,11 +161,29 @@ export async function runGenerateSubtitlesCommand(
    const mediaPath = options.mediaPath ? localMediaPath(options.mediaPath) : currentMedia;
    if (!mediaPath)
      throw new Error('Pass a local video file or open one in mpv before running generate-subs.');
+    // Capture audio and subtitle timing together before model setup can yield to playback changes.
+    const matchesCurrentMedia = currentMedia !== null && sameFile(currentMedia, mediaPath);
+    const tracks = matchesCurrentMedia
+      ? await deps
+          .mpvCommand(context.mpvSocketPath, ['get_property', 'track-list'], 1000)
+          .catch(() => null)
+      : null;
+    let references: SubtitleGenerationReference[] = [];
+    if (matchesCurrentMedia) {
+      const candidates = await readSubtitleGenerationReferences(tracks, (name) =>
+        deps.mpvCommand(context.mpvSocketPath, ['get_property', name], 1000),
+      );
+      const stillPlaying = await readMpvMedia(context.mpvSocketPath, deps.mpvCommand).catch(
+        () => null,
+      );
+      if (stillPlaying && sameFile(stillPlaying, mediaPath)) references = candidates;
+      else if (!options.mediaPath && options.audioStreamIndex === undefined)
+        throw new Error(
+          'The current mpv media changed or could not be verified while reading tracks. Run generate-subs again.',
+        );
+    }
    const audioStreamIndex =
-      options.audioStreamIndex ??
-      (!options.mediaPath
-        ? await readMpvAudioStream(context.mpvSocketPath, deps.mpvCommand)
-        : undefined);
+      options.audioStreamIndex ?? (!options.mediaPath ? selectedMpvAudioStream(tracks) : undefined);
    const onProgress = createGenerationProgressReporter(write);
    // Missing executables fail here, before any model download starts.
    requireSubtitleGenerationTools(await deps.resolveTools(config));
@@ -183,6 +202,7 @@ export async function runGenerateSubtitlesCommand(
      modelDirectory,
      mediaPath,
      audioStreamIndex,
+      references,
      outputPath: options.outputPath
        ? path.resolve(resolvePathMaybe(options.outputPath))
        : undefined,
@@ -89,7 +89,6 @@ export async function jellyfinApiRequest<T>(
  const url = `${session.serverUrl}${requestPath}`;
  const response = await fetch(url, {
    headers: {
-      'X-Emby-Token': session.accessToken,
      Authorization: `MediaBrowser Token="${session.accessToken}"`,
    },
  });
@@ -103,7 +102,7 @@ export async function jellyfinApiRequest<T>(
 }

 function itemPreviewUrl(session: JellyfinSessionConfig, id: string): string {
-  return `${session.serverUrl}/Items/${id}/Images/Primary?maxHeight=720&quality=85&api_key=${encodeURIComponent(session.accessToken)}`;
+  return `${session.serverUrl}/Items/${id}/Images/Primary?maxHeight=720&quality=85&ApiKey=${encodeURIComponent(session.accessToken)}`;
 }

 function jellyfinIconCacheDir(session: JellyfinSessionConfig): string {
@@ -228,7 +228,7 @@ export function pickLibrary(
    commandExists('chafa') && commandExists('curl')
      ? `
 id={1}
-url=${escapeShellSingle(session.serverUrl)}/Items/$id/Images/Primary?maxHeight=720\\&quality=85\\&api_key=${escapeShellSingle(session.accessToken)}
+url=${escapeShellSingle(session.serverUrl)}/Items/$id/Images/Primary?maxHeight=720\\&quality=85\\&ApiKey=${escapeShellSingle(session.accessToken)}
 curl -fsSL "$url" 2>/dev/null | chafa --format=symbols --symbols=vhalf+wide --size=${'${FZF_PREVIEW_COLUMNS}'}x${'${FZF_PREVIEW_LINES}'} - 2>/dev/null
 `.trim()
      : 'echo "Install curl + chafa for image preview"';
@@ -266,7 +266,7 @@ export function pickItem(
    commandExists('chafa') && commandExists('curl')
      ? `
 id={1}
-url=${escapeShellSingle(session.serverUrl)}/Items/$id/Images/Primary?maxHeight=720\\&quality=85\\&api_key=${escapeShellSingle(session.accessToken)}
+url=${escapeShellSingle(session.serverUrl)}/Items/$id/Images/Primary?maxHeight=720\\&quality=85\\&ApiKey=${escapeShellSingle(session.accessToken)}
 curl -fsSL "$url" 2>/dev/null | chafa --format=symbols --symbols=vhalf+wide --size=${'${FZF_PREVIEW_COLUMNS}'}x${'${FZF_PREVIEW_LINES}'} - 2>/dev/null
 `.trim()
      : 'echo "Install curl + chafa for image preview"';
@@ -304,7 +304,7 @@ export function pickGroup(
    commandExists('chafa') && commandExists('curl')
      ? `
 id={1}
-url=${escapeShellSingle(session.serverUrl)}/Items/$id/Images/Primary?maxHeight=720\\&quality=85\\&api_key=${escapeShellSingle(session.accessToken)}
+url=${escapeShellSingle(session.serverUrl)}/Items/$id/Images/Primary?maxHeight=720\\&quality=85\\&ApiKey=${escapeShellSingle(session.accessToken)}
 curl -fsSL "$url" 2>/dev/null | chafa --format=symbols --symbols=vhalf+wide --size=${'${FZF_PREVIEW_COLUMNS}'}x${'${FZF_PREVIEW_LINES}'} - 2>/dev/null
 `.trim()
      : 'echo "Install curl + chafa for image preview"';
@@ -31,6 +31,7 @@ const SYNC_SCHEMA_OBJECTS = [
  'imm_lifetime_applied_sessions',
  'imm_stats_excluded_words',
  'idx_anime_normalized_title',
+  'idx_anime_namespace_title',
  'idx_anime_anilist_id',
  'idx_videos_anime_id',
  'idx_sessions_video_started',
@@ -12,7 +12,7 @@ export const IMMERSION_DB_FIXTURE_DDL = `
  );
  CREATE TABLE imm_anime(
    anime_id INTEGER PRIMARY KEY AUTOINCREMENT,
-    normalized_title_key TEXT NOT NULL UNIQUE,
+    normalized_title_key TEXT NOT NULL,
    canonical_title TEXT NOT NULL,
    anilist_id INTEGER UNIQUE,
    title_romaji TEXT,
@@ -20,10 +20,14 @@ export const IMMERSION_DB_FIXTURE_DDL = `
    title_native TEXT,
    episodes_total INTEGER,
    description TEXT,
+    media_kind TEXT NOT NULL DEFAULT 'anime',
+    tmdb_id INTEGER,
+    tmdb_type TEXT,
    metadata_json TEXT,
    CREATED_DATE TEXT,
    LAST_UPDATE_DATE TEXT
  );
+  CREATE UNIQUE INDEX idx_anime_namespace_title ON imm_anime((media_kind = 'youtube'), normalized_title_key);
  CREATE TABLE imm_videos(
    video_id INTEGER PRIMARY KEY AUTOINCREMENT,
    video_key TEXT NOT NULL UNIQUE,
@@ -2,7 +2,7 @@
  "name": "subminer",
  "productName": "SubMiner",
  "desktopName": "SubMiner.desktop",
-  "version": "0.19.6",
+  "version": "0.20.0-beta.1",
  "description": "All-in-one sentence mining overlay with AnkiConnect and dictionary integration",
  "packageManager": "bun@1.3.5",
  "main": "dist/main-entry.js",
@@ -51,7 +51,7 @@
    "test:docs:kb": "bun test scripts/docs-knowledge-base.test.ts",
    "test:plugin:src": "lua scripts/test-plugin-lua-compat.lua && lua scripts/test-plugin-start-gate.lua && lua scripts/test-plugin-process-start-retries.lua && lua scripts/test-plugin-restart-feedback.lua && lua scripts/test-plugin-session-bindings.lua && lua scripts/test-plugin-binary-windows.lua",
    "test:launcher:smoke:src": "bun test launcher/smoke.e2e.test.ts",
-    "test:smoke:dist": "bun scripts/run-test-lane.mjs bun-src-full",
+    "test:smoke:dist": "env ELECTRON_RUN_AS_NODE=1 electron scripts/compiled-runtime-smoke.mjs",
    "test:subtitle:src": "bun test src/core/services/subsync.test.ts src/subsync/utils.test.ts",
    "test:immersion:sqlite:src": "bun test src/core/services/immersion-tracker-service.test.ts src/core/services/immersion-tracker/storage-session.test.ts",
    "test:immersion:sqlite:dist": "bun test dist/core/services/immersion-tracker-service.test.js dist/core/services/immersion-tracker/storage-session.test.js",
@@ -63,7 +63,7 @@
    "test:scripts": "bun scripts/run-test-lane.mjs scripts",
    "test:stats": "bun scripts/run-test-lane.mjs stats",
    "test:env": "bun run test:launcher:smoke:src && bun run test:plugin:src && bun run test:immersion:sqlite:src",
-    "test:runtime:compat": "bun run tsc && bun scripts/run-test-lane.mjs bun-src-full",
+    "test:runtime:compat": "bun run test:smoke:dist",
    "test": "bun run test:fast",
    "test:config": "bun scripts/run-test-lane.mjs config",
    "test:launcher": "bun scripts/run-test-lane.mjs launcher && bun run test:plugin:src",
@@ -85,9 +85,9 @@
  },
  "overrides": {
    "@xmldom/xmldom": "0.8.15",
-    "app-builder-lib": "26.15.3",
+    "app-builder-lib": "26.16.1",
    "brace-expansion": "5.0.9",
-    "electron-builder-squirrel-windows": "26.15.3",
+    "electron-builder-squirrel-windows": "26.16.1",
    "fast-uri": "3.1.6",
    "form-data": "4.0.6",
    "ip-address": "10.2.0",
@@ -129,7 +129,7 @@
    "@types/node": "^24.10.0",
    "@types/ws": "^8.18.1",
    "electron": "42.6.0",
-    "electron-builder": "26.15.3",
+    "electron-builder": "26.16.1",
    "esbuild": "^0.25.12",
    "eslint": "^10.8.0",
    "prettier": "^3.8.1",
@@ -1,68 +1,107 @@
 > This is a prerelease build for testing. Stable changelog and docs-site updates remain pending until the final stable release.

-<!-- prerelease-base-version: 0.19.4 -->
+<!-- prerelease-version: 0.20.0-beta.1 -->

 ## Highlights
 ### Added

- Library Merge & Reassignment
-  - Duplicate library cards for the same show can be combined: select entries in "Select" mode and use "Merge Selected" to combine their sessions, mined cards, and watch time onto one card.
-  - Episodes can be moved to a different entry with a per-episode "→" button, fixing stray files that split off their own entry; manual assignments now survive later filename parsing, Jellyfin refreshes, and season repair.
-  - Exact AniList matches with compatible seasons merge automatically, while likely (fuzzy) matches surface as a dismissible "Possible duplicate" suggestion instead of merging silently.
+- **Japanese Subtitle Generation**:
+  - Generate Japanese subtitles locally with whisper.cpp, right from a modal (`Ctrl+Shift+G`), the subtitle sidebar's generation button when no subtitles are loaded, or `subminer generate-subs`, with progress, cancellation, and automatic loading into mpv when it's done.
+  - Pick and download an official multilingual Whisper model in-app (including smaller quantized variants), or point Settings at one you already have. SubMiner recommends `large-v3-turbo` when CUDA is available and `small` otherwise, and tells you up front if `whisper-cli`, `ffmpeg`, or `ffprobe` can't be found.
+  - An optional "Focus on spoken dialogue" mode uses a Silero VAD model to keep quiet or music-covered dialogue that would otherwise get dropped.
+  - Long passages split near natural speech pauses, guided by an existing subtitle track when one is loaded, giving tighter timing and fewer repeated-word glitches.

- Duplicate Line Cleanup
-  - The Vocabulary tab's new **Duplicates** button scans a chosen time window for the repeated-line bursts described under Fixed below and collapses each burst to a single line once confirmed.
-  - A matching `subminer stats cleanup --duplicate-lines` command (with `--dry-run` and `--lookback-days <n>`) is available from the terminal.
-  - Only the affected subtitle lines and the vocabulary counts they inflated are touched; watch time and lines-seen totals are left as recorded.
+- **Media Timing Review Frame Picker**:
+  - The screenshot used for a mined card can now be chosen independently of the audio clip, with its own live preview, time slider, and frame-by-frame stepping.
+  - Works for local video and for seekable remote streams like Jellyfin.
+
+- **Overlay Keybinding Pickup**: The overlay now recognizes your mpv keybindings (from mpv's defaults, `input.conf`, and loaded scripts) as long as they don't conflict with SubMiner's own controls. Picked-up bindings work for the session but won't show up in the help menu.
+
+- **Subtitle Sidebar Selection & Copy**: You can now select dialogue across multiple subtitle sidebar rows and copy it, without timestamps, using Ctrl/Cmd+C or the Copy button, without seeking or mining a card.
+
+- **Jimaku Live Action Search**: The Jimaku modal has separate Anime and Live Action tabs (switch with Arrow Left/Right) so you can search Jimaku's live-action subtitle catalogue directly.
+
+- **Live-Action TMDB Library**:
+  - Live-action dramas and movies in the stats Library now get posters, synopses, and titles from TMDB.
+  - Titles AniList can't match are looked up on TMDB automatically when the parsed filename matches a title exactly; otherwise use the new **Link to TMDB** action. Entries linked to the same TMDB title merge into one card, and the Library kind selector gained a Live Action option.
+  - Release builds already include a TMDB key; if you run from source, set `tmdb.apiKey` (or `tmdb.apiKeyCommand`) yourself.
+
+- **YouTube Library Kind**:
+  - YouTube channels are now their own Library media kind, with new All Titles, Anime, and YouTube filters. Existing channel entries migrate automatically with viewing history and manual video assignments intact.
+  - Channels stay out of AniList matching, season repair, and duplicate recommendations, and can't be merged or moved into an anime entry.
+
+### Changed
+
+- **Bundled Bun Runtime**: Every SubMiner launcher, installed or downloaded, now runs on the Bun runtime bundled with the app instead of a system-wide Bun install. Recognized legacy launchers migrate automatically, Windows users get a new `subminer.cmd` download, and first-run setup now shows a single optional launcher control with runtime repair guidance only when something actually needs it.
+
+- **Compressed Incremental Sync**: Cross-machine sync between compatible macOS/Linux machines now transfers only what changed, compressed, using a cached snapshot from the last sync to cut traffic further. Machines without a compatible rsync (including Windows) fall back to compressed scp automatically, older peers keep working, and transfers now time out after 30 minutes instead of hanging indefinitely.
+
+- **Smaller Install Size**: Installers and the unpacked app are smaller after dropping demo media, source maps, TypeScript sources, test fixtures, and unused binaries, and sharing one Japanese UI font across windows. Release builds now publish a package-size comparison against the previous release.
+
+- **Stats Server Request Safety**: The stats server, including the in-app stats overlay which now loads through it, only accepts requests from the local machine and requires a JSON content type for anything that changes data. If you were exposing the dashboard through a reverse proxy or Tailscale Serve, that's no longer supported, and any script posting to the stats API needs to send `Content-Type: application/json`.
+
+- **Yomitan Updated**: Bundled Yomitan is updated to upstream 26.9.8, adding historical Japanese kana transformations and Ukrainian language support, plus improvements to Anki duplicate search and audio retrieval.

 ### Fixed

- Subtitle Duplication from Karaoke & Animated Signs
-  - Typeset ASS karaoke and animated signs no longer flood the overlay, subtitle sidebar, immersion history, mined cards, or stats with repeated glyph fragments or per-frame duplicates; the complete authored line is recovered instead, without merging genuinely repeated dialogue or separately positioned signs.
-  - Fragmented karaoke now preserves the spaces the author placed between words instead of joining them together, and lyric transitions (including seeking into the middle of a line) resolve to the clean line instead of a stray entrance or exit frame.
-  - The secondary overlay shares the same deduplication logic as the primary overlay, including collapsing lines that differ only by whitespace or trailing punctuation, and sidebar navigation moves between clean lyric lines while keeping the right line selected.
+- **Jellyfin**:
+  - Playback, subtitles, artwork, and remote control now authenticate with an `ApiKey` parameter instead of legacy headers, so Jellyfin 12 works correctly even with legacy authorization disabled.
+  - "Play on SubMiner" no longer silently drops the connection after about a minute on Jellyfin 12.
+  - The "now playing" bar clears when you close or finish a cast video instead of running to the end of the episode.
+  - Anki cards mined from Jellyfin now get the real episode title in the misc info field instead of "Unknown media".
+  - Jellyfin streams no longer leak URL-derived titles or credential-bearing URLs into metadata, Anki fields, Discord presence, stats, or AniList lookups; previously cached data that had credentials in it is cleaned up automatically.

- Anki Media Generation
-  - Sentence-audio generation no longer times out on slow network-mounted video files with many subtitle and font streams, and a failed extraction now reports a clear error instead of a raw `ENOENT`.
-  - Mined audio and animated AVIF clips now capture the subtitle line you actually mined, instead of whatever line happened to be on screen once slow audio extraction finished.
+- **Anki & Mining**:
+  - Word audio now reads from its own configured field (`ankiConnect.fields.wordAudio`) instead of the sentence-audio field, fixing animated word images that started moving immediately instead of on demand.
+  - Setting `ankiConnect.media.maxMediaDuration` to `0` for unlimited duration now also applies when mining from the stats dashboard, matching overlay mining.
+  - Closing the overlay while a media timing review is still loading now properly cancels setup, restores playback if the review had paused it, and cleans up the hidden preview player.

- Character Dictionary Performance & Notifications
-  - Character dictionary generation, merged rebuilds, and imports no longer freeze the app on large dictionaries, and cached results (including character portraits) are reused across launches instead of regenerating everything every time.
-  - Portraits also now display correctly if their cache finishes loading after subtitles have already started showing.
-  - Desktop progress notifications, including on Linux AppImage installs, now update in place instead of flickering closed and reopening.
+- **Settings**:
+  - AnkiConnect, Kiku, and Senren settings are now validated before use, with a warning and a safe default for anything invalid instead of a bad value reaching runtime.
+  - Settings marked as applying live now correctly avoid showing a restart warning, and mixed saves apply the live parts immediately while listing only the sections that actually need a restart.

- Overlay Reliability
-  - Overlay modals (settings, stats, etc.) now open promptly on the first shortcut press, including on repeated sessions on Windows, and appear above fullscreen mpv on macOS instead of switching Spaces or opening off-screen.
-  - The macOS window-tracking helper is now built for macOS 12.0+, so the overlay attaches to mpv on older systems like Ventura instead of crashing and getting stuck on "Overlay loading."
-  - The overlay no longer gets stuck on "Overlay loading" indefinitely if mpv's connection stalls; it now retries and shows an actionable error after 30 seconds.
-  - Fixed native Wayland drag-and-drop from file managers like Thunar, and fixed system-wide mouse lag on Windows caused by the overlay's click-through handling.
+- **Overlay**:
+  - Clicking a subtitle sidebar cue no longer leaves Space bound to seeking back to it; Enter still seeks the focused cue, and Space keeps whatever playback action you've configured.
+  - Hyprland recovery dialogs now stay above SubMiner windows instead of being covered by overlay placement updates.
+  - Fixed a rare case on Linux where a delayed window-close callback could reopen the overlay after it was torn down.

- Stats Dashboard
-  - Deletes, library merges, video moves, and AniList reassignments no longer freeze the stats dashboard or rebuild lifetime totals from scratch; large deletes that used to take minutes now finish in milliseconds.
-  - Vocabulary totals and charts now count all tracked vocabulary instead of just the first page, and new-word history uses corrected daily rollups.
-  - Calendar labels respect time zones west of UTC, and vocabulary cards refresh automatically after editing the word exclusion list (with a Retry option if a load fails).
+- **Stats**:
+  - Malformed or partly invalid resource IDs are now rejected before they can affect library mutations or cover-art backfills.
+  - Stats server port conflicts now surface as a status notification instead of crashing SubMiner, and startup/shutdown are more robust: concurrent startup requests share one attempt, stopping a background instance no longer disconnects an open dashboard, and shutdown no longer waits indefinitely on active requests.

- Linux Launcher Thumbnails
-  - Fixed missing MKV thumbnails in the Linux rofi picker when the system thumbnailer only registers legacy Matroska MIME aliases.
+- **Subtitle Sidebar Gap Follow**: The subtitle sidebar now stays near actual playback position during gaps in files where a cue starts at time zero.
+
+- **First Launch on macOS**: Fixed first launch exiting immediately when the SubMiner config directory didn't exist yet.

 ## What's Changed

- feat(stats): add library entry merge and episode move by @ksyasuda in #190
- fix(stats): stop counting duplicate typeset subtitle lines by @ksyasuda in #191
- fix(media): tolerate slow MKV audio extraction by @ksyasuda in #195
- fix(stats): subtract lifetime totals incrementally on delete by @ksyasuda in #196
- fix(anki): snapshot mining media clip timing by @ksyasuda in #197
- fix(notifications): replace Linux progress updates in place by @ksyasuda in #198
- fix(overlay): support native Wayland file drag-and-drop by @ksyasuda in #199
- fix(overlay): keep macOS modal windows on fullscreen Spaces by @ksyasuda in #200
- fix(overlay): prevent Windows mouse lag during click-through tracking by @ksyasuda in #201
- fix(stats): report complete vocabulary totals and new-word history by @ksyasuda in #202
- fix(mpv): recover from stalled IPC connects by @ksyasuda in #204
- fix(dictionary): prevent freezes and restore AppImage notifications by @ksyasuda in #205
- fix(subtitles): recover canonical lines from ASS animation by @ksyasuda in #207
- fix(overlay): deduplicate secondary subtitle rendering by @ksyasuda in #208
- fix(launcher): restore Matroska thumbnails in Linux rofi picker by @ksyasuda in #210
- fix(character-dictionary): cache completed MeCab refreshes by @ksyasuda in #212
+- feat(sidebar): add dialogue selection and copying by @ksyasuda in #238
+- feat(subtitles): add local Japanese subtitle generation by @ksyasuda in #240
+- perf(stats): use compressed incremental snapshot transfers by @ksyasuda in #241
+- fix(startup): create config directory before singleton lock by @ksyasuda in #242
+- feat(launcher): bundle Bun and use it across all launchers by @ksyasuda in #243
+- build(release): reduce package size and report release sizes by @ksyasuda in #244
+- fix(overlay): keep Hyprland recovery dialogs above overlays by @ksyasuda in #245
+- feat(overlay): discover unclaimed mpv key bindings by @ksyasuda in #246
+- fix(sidebar): preserve Space playback after cue seeking by @ksyasuda in #247
+- fix(jellyfin): fix jellyfin media metadata by @ksyasuda in #250
+- feat(jimaku): add live-action subtitle search by @ksyasuda in #251
+- feat(stats): add TMDB metadata for live-action dramas in the Library by @ksyasuda in #252
+- feat(stats): separate YouTube channels in the Library by @ksyasuda in #253
+- feat(mining): add a screenshot frame picker to media review by @aalhendi in #254
+- fix(config): align live save feedback with hot reload policy by @ksyasuda in #255
+- fix(anki): separate word audio mapping for animation sync by @ksyasuda in #256
+- fix(config): validate AnkiConnect and field grouping settings by @ksyasuda in #257
+- fix(anki): honor unlimited duration in stats mining by @ksyasuda in #258
+- fix(stats): reject malformed resource IDs before mutations by @ksyasuda in #259
+- fix(stats): harden server lifecycle and verify compiled runtime by @ksyasuda in #261
+- fix(overlay): cancel pending window transitions and timing reviews by @ksyasuda in #262
+- fix(stats): restrict local requests and serve the dashboard over HTTP by @ksyasuda in #263
+- fix(jellyfin): support modern authentication by @ksyasuda in #264
+
+## New Contributors
+
+- @aalhendi made their first contribution in #254

 ## Installation

@@ -0,0 +1,29 @@
+import fs from 'node:fs';
+import path from 'node:path';
+
+/**
+ * Release builds carry a project-owned TMDB key so live-action lookups work
+ * without user setup. The key is injected from the SUBMINER_TMDB_API_KEY
+ * environment variable at build time (a GitHub Actions secret in CI) and never
+ * lives in the repository. The runtime reader is
+ * src/core/services/tmdb/bundled-api-key.ts; keep the file name in sync.
+ */
+export const BUNDLED_INTEGRATION_KEYS_FILENAME = 'bundled-integration-keys.json';
+export const TMDB_API_KEY_ENV = 'SUBMINER_TMDB_API_KEY';
+
+/**
+ * Write the bundled keys file into `distDir`, or remove a stale one when no
+ * key is present so a keyless build never ships an older key by accident.
+ * Returns the names of the keys staged.
+ */
+export function stageBundledIntegrationKeys(distDir, env = process.env) {
+  const outputPath = path.join(distDir, BUNDLED_INTEGRATION_KEYS_FILENAME);
+  const tmdbApiKey = env[TMDB_API_KEY_ENV]?.trim() ?? '';
+  if (!tmdbApiKey) {
+    fs.rmSync(outputPath, { force: true });
+    return [];
+  }
+  fs.mkdirSync(distDir, { recursive: true });
+  fs.writeFileSync(outputPath, `${JSON.stringify({ tmdbApiKey })}\n`, { mode: 0o644 });
+  return ['tmdb'];
+}
@@ -0,0 +1,39 @@
+import assert from 'node:assert/strict';
+import fs from 'node:fs';
+import os from 'node:os';
+import path from 'node:path';
+import test from 'node:test';
+import {
+  BUNDLED_INTEGRATION_KEYS_FILENAME,
+  stageBundledIntegrationKeys,
+} from './bundled-integration-keys.mjs';
+
+function withDistDir(work: (distDir: string) => void): void {
+  const distDir = fs.mkdtempSync(path.join(os.tmpdir(), 'subminer-bundled-keys-'));
+  try {
+    work(distDir);
+  } finally {
+    fs.rmSync(distDir, { recursive: true, force: true });
+  }
+}
+
+test('stageBundledIntegrationKeys writes the TMDB key from the environment', () => {
+  withDistDir((distDir) => {
+    const staged = stageBundledIntegrationKeys(distDir, { SUBMINER_TMDB_API_KEY: ' abc123 ' });
+    assert.deepEqual(staged, ['tmdb']);
+    const written = JSON.parse(
+      fs.readFileSync(path.join(distDir, BUNDLED_INTEGRATION_KEYS_FILENAME), 'utf8'),
+    );
+    assert.deepEqual(written, { tmdbApiKey: 'abc123' });
+  });
+});
+
+test('stageBundledIntegrationKeys removes a stale file when the variable is unset', () => {
+  withDistDir((distDir) => {
+    const outputPath = path.join(distDir, BUNDLED_INTEGRATION_KEYS_FILENAME);
+    fs.writeFileSync(outputPath, '{"tmdbApiKey":"old"}');
+    assert.deepEqual(stageBundledIntegrationKeys(distDir, {}), []);
+    assert.equal(fs.existsSync(outputPath), false);
+    assert.deepEqual(stageBundledIntegrationKeys(distDir, { SUBMINER_TMDB_API_KEY: '  ' }), []);
+  });
+});
@@ -0,0 +1,345 @@
+import assert from 'node:assert/strict';
+import { spawn } from 'node:child_process';
+import { existsSync, mkdtempSync, readFileSync, rmSync, statSync, writeFileSync } from 'node:fs';
+import { createServer } from 'node:net';
+import { request as httpRequest } from 'node:http';
+import { tmpdir } from 'node:os';
+import { dirname, join, resolve } from 'node:path';
+import { fileURLToPath } from 'node:url';
+
+const START_TIMEOUT_MS = 15_000;
+const STOP_TIMEOUT_MS = 5_000;
+const POLL_INTERVAL_MS = 25;
+const MAX_CHILD_OUTPUT_BYTES = 64 * 1024;
+
+const scriptDir = dirname(fileURLToPath(import.meta.url));
+const repoRootArgIndex = process.argv.indexOf('--repo-root');
+const repoRoot = resolve(
+  repoRootArgIndex === -1 ? join(scriptDir, '..') : (process.argv[repoRootArgIndex + 1] ?? ''),
+);
+const paths = {
+  mainEntry: join(repoRoot, 'dist', 'main-entry.js'),
+  daemonEntry: join(repoRoot, 'dist', 'stats-daemon-runner.js'),
+  statsServer: join(repoRoot, 'dist', 'core', 'services', 'stats-server.js'),
+  tracker: join(repoRoot, 'dist', 'core', 'services', 'immersion-tracker-service.js'),
+  statsIndex: join(repoRoot, 'stats', 'dist', 'index.html'),
+};
+
+function requireCompiledArtifacts() {
+  const missing = Object.values(paths).filter((artifactPath) => !existsSync(artifactPath));
+  if (missing.length > 0) {
+    throw new Error(
+      `Compiled runtime artifacts are missing. Run \`bun run build\` before this check:\n${missing
+        .map((artifactPath) => `  - ${artifactPath}`)
+        .join('\n')}`,
+    );
+  }
+}
+
+function requireElectronNodeRuntime() {
+  if (!process.versions.electron || process.env.ELECTRON_RUN_AS_NODE !== '1') {
+    throw new Error(
+      'This check must run with Electron in Node mode. Use `bun run test:smoke:dist`.',
+    );
+  }
+}
+
+function delay(ms) {
+  return new Promise((resolvePromise) => setTimeout(resolvePromise, ms));
+}
+
+function listen(server, port = 0) {
+  return new Promise((resolvePromise, reject) => {
+    const onError = (error) => {
+      server.off('listening', onListening);
+      reject(error);
+    };
+    const onListening = () => {
+      server.off('error', onError);
+      const address = server.address();
+      if (!address || typeof address === 'string') {
+        reject(new Error('Could not resolve the stats smoke port.'));
+        return;
+      }
+      resolvePromise(address.port);
+    };
+    server.once('error', onError);
+    server.once('listening', onListening);
+    server.listen(port, '127.0.0.1');
+  });
+}
+
+function closeServer(server) {
+  return new Promise((resolvePromise, reject) => {
+    server.close((error) => {
+      if (error) {
+        reject(error);
+        return;
+      }
+      resolvePromise();
+    });
+  });
+}
+
+async function findAvailablePort() {
+  const reservation = createServer();
+  const port = await listen(reservation);
+  await closeServer(reservation);
+  return port;
+}
+
+function writeSmokeConfig(userDataPath, port) {
+  writeFileSync(
+    join(userDataPath, 'config.json'),
+    `${JSON.stringify({ stats: { serverPort: port } })}\n`,
+  );
+}
+
+function spawnDaemon(userDataPath, responsePath) {
+  const child = spawn(
+    process.execPath,
+    [
+      paths.daemonEntry,
+      '--stats-user-data-path',
+      userDataPath,
+      '--stats-response-path',
+      responsePath,
+    ],
+    {
+      cwd: repoRoot,
+      env: { ...process.env, ELECTRON_RUN_AS_NODE: '1' },
+      stdio: ['ignore', 'pipe', 'pipe'],
+    },
+  );
+  let output = '';
+  const appendOutput = (chunk) => {
+    if (output.length >= MAX_CHILD_OUTPUT_BYTES) return;
+    output += String(chunk);
+    if (output.length > MAX_CHILD_OUTPUT_BYTES) {
+      output = `${output.slice(0, MAX_CHILD_OUTPUT_BYTES)}\n[child output truncated]\n`;
+    }
+  };
+  child.stdout.on('data', appendOutput);
+  child.stderr.on('data', appendOutput);
+  let exited = false;
+  const exit = new Promise((resolvePromise) => {
+    child.once('error', (error) => {
+      exited = true;
+      resolvePromise({ code: null, signal: null, error });
+    });
+    child.once('exit', (code, signal) => {
+      exited = true;
+      resolvePromise({ code, signal, error: null });
+    });
+  });
+  return { child, exit, hasExited: () => exited, getOutput: () => output };
+}
+
+async function waitForResponse(responsePath, daemon) {
+  const deadline = Date.now() + START_TIMEOUT_MS;
+  while (Date.now() < deadline) {
+    if (existsSync(responsePath)) {
+      try {
+        return JSON.parse(readFileSync(responsePath, 'utf8'));
+      } catch {
+        // The daemon may still be finishing the response file write.
+      }
+    }
+    if (daemon.hasExited()) {
+      const result = await daemon.exit;
+      throw new Error(
+        `Stats daemon exited before writing a startup response (${formatExit(result)}).\n${daemon.getOutput()}`,
+      );
+    }
+    await delay(POLL_INTERVAL_MS);
+  }
+  throw new Error(`Timed out waiting for stats daemon startup.\n${daemon.getOutput()}`);
+}
+
+function formatExit(result) {
+  if (result.error) return result.error.message;
+  if (result.signal) return `signal ${result.signal}`;
+  return `exit code ${result.code}`;
+}
+
+async function waitForExit(daemon, timeoutMs) {
+  return await Promise.race([daemon.exit, delay(timeoutMs).then(() => null)]);
+}
+
+async function stopDaemon(daemon) {
+  if (daemon.hasExited()) {
+    return await daemon.exit;
+  }
+  daemon.child.kill('SIGTERM');
+  const result = await waitForExit(daemon, STOP_TIMEOUT_MS);
+  if (result) return result;
+  daemon.child.kill('SIGKILL');
+  await daemon.exit;
+  throw new Error(`Stats daemon did not stop after SIGTERM.\n${daemon.getOutput()}`);
+}
+
+async function assertPortCanBind(port) {
+  const server = createServer();
+  try {
+    await listen(server, port);
+  } finally {
+    if (server.listening) await closeServer(server);
+  }
+}
+
+async function fetchOverview(url, daemon) {
+  const deadline = Date.now() + START_TIMEOUT_MS;
+  let lastError = null;
+  while (Date.now() < deadline) {
+    try {
+      return await fetch(`${url}/api/stats/overview`, {
+        signal: AbortSignal.timeout(START_TIMEOUT_MS),
+      });
+    } catch (error) {
+      lastError = error;
+    }
+    if (daemon.hasExited()) {
+      const result = await daemon.exit;
+      throw new Error(
+        `Stats daemon exited before accepting HTTP requests (${formatExit(result)}).\n${daemon.getOutput()}`,
+      );
+    }
+    await delay(POLL_INTERVAL_MS);
+  }
+  throw new Error(
+    `Timed out waiting for the stats HTTP server: ${lastError instanceof Error ? lastError.message : String(lastError)}\n${daemon.getOutput()}`,
+  );
+}
+
+async function runHealthyStartup(userDataPath, port, responseName) {
+  writeSmokeConfig(userDataPath, port);
+  const responsePath = join(userDataPath, responseName);
+  const statePath = join(userDataPath, 'stats-daemon.json');
+  const databasePath = join(userDataPath, 'immersion.sqlite');
+  const daemon = spawnDaemon(userDataPath, responsePath);
+  let shutdownResult;
+  let unfinishedRequest;
+
+  try {
+    const startup = await waitForResponse(responsePath, daemon);
+    assert.deepEqual(startup, { ok: true, url: `http://127.0.0.1:${port}` });
+
+    const response = await fetchOverview(startup.url, daemon);
+    assert.equal(response.status, 200);
+    assert.match(response.headers.get('content-type') ?? '', /^application\/json\b/);
+    const overview = await response.json();
+    assert.equal(typeof overview, 'object');
+    assert.ok(overview !== null);
+    assert.ok(Array.isArray(overview.sessions));
+    assert.ok(Array.isArray(overview.rollups));
+    assert.equal(typeof overview.hints, 'object');
+
+    const dashboard = await fetch(`${startup.url}/?overlay=1`, {
+      signal: AbortSignal.timeout(START_TIMEOUT_MS),
+    });
+    assert.equal(dashboard.status, 200);
+    assert.match(dashboard.headers.get('content-type') ?? '', /^text\/html\b/);
+    assert.match(await dashboard.text(), /id="root"/);
+
+    assert.ok(existsSync(databasePath), 'The compiled tracker did not create its SQLite database.');
+    assert.ok(
+      statSync(databasePath).size > 0,
+      'The compiled tracker created an empty SQLite file.',
+    );
+
+    // Leave a real request body unfinished so shutdown must bound its drain wait.
+    unfinishedRequest = httpRequest(new URL('/api/stats/anki/notesInfo', startup.url), {
+      method: 'POST',
+      headers: {
+        'Content-Type': 'application/json',
+        'Content-Length': '1000',
+        Expect: '100-continue',
+      },
+      signal: AbortSignal.timeout(START_TIMEOUT_MS),
+    });
+    await new Promise((resolvePromise, reject) => {
+      unfinishedRequest.on('error', reject);
+      unfinishedRequest.once('continue', () => {
+        unfinishedRequest.write('{');
+        resolvePromise();
+      });
+      unfinishedRequest.flushHeaders();
+    });
+  } finally {
+    try {
+      shutdownResult = await stopDaemon(daemon);
+    } finally {
+      unfinishedRequest?.destroy();
+    }
+  }
+
+  assert.equal(shutdownResult.code, 0, `Stats daemon shutdown failed.\n${daemon.getOutput()}`);
+  assert.equal(existsSync(statePath), false, 'Stats daemon state remained after shutdown.');
+  await assertPortCanBind(port);
+}
+
+async function runConflictRecovery() {
+  const userDataPath = mkdtempSync(join(tmpdir(), 'subminer-compiled-conflict-'));
+  const reservation = createServer();
+  const port = await listen(reservation);
+  const responsePath = join(userDataPath, 'conflict-response.json');
+  const statePath = join(userDataPath, 'stats-daemon.json');
+  writeSmokeConfig(userDataPath, port);
+  const daemon = spawnDaemon(userDataPath, responsePath);
+  const failures = [];
+
+  try {
+    const response = await waitForResponse(responsePath, daemon);
+    if (response.ok !== false) {
+      failures.push('The stats daemon reported success while its configured port was occupied.');
+    }
+    const result = await waitForExit(daemon, STOP_TIMEOUT_MS);
+    if (!result) {
+      failures.push('The stats daemon did not exit after its configured port failed to bind.');
+    } else if (result.code === 0) {
+      failures.push(
+        'The stats daemon exited successfully after its configured port failed to bind.',
+      );
+    }
+    if (existsSync(statePath)) {
+      failures.push(
+        'The stats daemon left ownership state behind after its configured port failed to bind.',
+      );
+    }
+  } finally {
+    await closeServer(reservation);
+    if (!daemon.hasExited()) {
+      await stopDaemon(daemon);
+    }
+  }
+
+  try {
+    await runHealthyStartup(userDataPath, port, 'recovery-response.json');
+  } finally {
+    rmSync(userDataPath, { recursive: true, force: true });
+  }
+
+  assert.deepEqual(failures, []);
+}
+
+async function main() {
+  requireCompiledArtifacts();
+  requireElectronNodeRuntime();
+
+  const userDataPath = mkdtempSync(join(tmpdir(), 'subminer-compiled-runtime-'));
+  try {
+    await runHealthyStartup(userDataPath, await findAvailablePort(), 'startup-response.json');
+  } finally {
+    rmSync(userDataPath, { recursive: true, force: true });
+  }
+  await runConflictRecovery();
+
+  process.stdout.write(
+    `Compiled runtime smoke passed with Electron ${process.versions.electron}, Node ${process.versions.node}, HTTP, and native SQLite.\n`,
+  );
+}
+
+main().catch((error) => {
+  process.stderr.write(`${error instanceof Error ? error.stack : String(error)}\n`);
+  process.exitCode = 1;
+});
@@ -0,0 +1,28 @@
+import assert from 'node:assert/strict';
+import { spawnSync } from 'node:child_process';
+import fs from 'node:fs';
+import os from 'node:os';
+import path from 'node:path';
+import test from 'node:test';
+
+test('compiled runtime smoke fails clearly when build artifacts are missing', () => {
+  const emptyRepo = fs.mkdtempSync(path.join(os.tmpdir(), 'subminer-compiled-missing-'));
+  try {
+    const result = spawnSync(
+      process.execPath,
+      ['scripts/compiled-runtime-smoke.mjs', '--repo-root', emptyRepo],
+      {
+        cwd: path.resolve(import.meta.dir, '..'),
+        encoding: 'utf8',
+      },
+    );
+
+    assert.equal(result.status, 1);
+    assert.match(result.stderr, /Compiled runtime artifacts are missing/);
+    assert.match(result.stderr, /dist\/main-entry\.js/);
+    assert.match(result.stderr, /dist\/stats-daemon-runner\.js/);
+    assert.doesNotMatch(result.stderr, /must run with Electron/);
+  } finally {
+    fs.rmSync(emptyRepo, { recursive: true, force: true });
+  }
+});
@@ -51,10 +51,13 @@ function listFiles(root, prefix = '') {
  });
 }

+// asar resolves lookups with the platform separator, so stat with the listed
+// native path and only normalize the reported name.
 function listAppFiles(archive) {
  return asar.listPackage(archive).flatMap((entry) => {
-    const name = entry.replaceAll('\\', '/').replace(/^\//, '');
-    const stat = asar.statFile(archive, name);
+    const native = entry.replace(/^[\\/]/, '');
+    const stat = asar.statFile(archive, native);
+    const name = native.replaceAll('\\', '/');
    return 'size' in stat ? [{ path: name, bytes: stat.size }] : [];
  });
 }
@@ -129,7 +132,7 @@ function verifyContents(archive, resources, platform, arch) {
    assert(!name.path.startsWith('minecard'), `Demo media shipped: ${name.path}`);
  }
  for (const ui of ['renderer', 'settings', 'syncui']) {
-    const css = asar.extractFile(archive, `dist/${ui}/style.css`).toString();
+    const css = asar.extractFile(archive, path.join('dist', ui, 'style.css')).toString();
    assert(css.includes('../fonts/MPLUS1[wght].ttf'), `Shared font missing from ${ui} CSS`);
  }
  return entries;
@@ -134,10 +134,12 @@ test('archive inventory handles native files without counting them twice on disk
    mkdirSync(output);
    writeFileSync(path.join(input, 'main.js'), 'hello');
    writeFileSync(path.join(input, 'native.node'), 'native');
+    mkdirSync(path.join(input, 'dist', 'ai'), { recursive: true });
+    writeFileSync(path.join(input, 'dist', 'ai', 'client.js'), 'nested');
    const archive = path.join(output, 'app.asar');
    await createPackageFromStreams(
      archive,
-      ['main.js', 'native.node'].map((name) => ({
+      ['main.js', 'native.node', 'dist/ai/client.js'].map((name) => ({
        path: name,
        type: 'file',
        unpacked: name.endsWith('.node'),
@@ -148,6 +150,7 @@ test('archive inventory handles native files without counting them twice on disk
    assert.deepEqual(listAppFiles(archive), [
      { path: 'main.js', bytes: 5 },
      { path: 'native.node', bytes: 6 },
+      { path: 'dist/ai/client.js', bytes: 6 },
    ]);
    assert.equal(
      listFiles(output).reduce((sum: number, entry: { bytes: number }) => sum + entry.bytes, 0),
@@ -3,6 +3,7 @@ import os from 'node:os';
 import path from 'node:path';
 import { execFileSync } from 'node:child_process';
 import { fileURLToPath } from 'node:url';
+import { stageBundledIntegrationKeys, TMDB_API_KEY_ENV } from './bundled-integration-keys.mjs';

 const scriptDir = path.dirname(fileURLToPath(import.meta.url));
 const repoRoot = path.resolve(scriptDir, '..');
@@ -97,6 +98,17 @@ function buildMacosHelper() {
  }
 }

+// Only the key names are logged, never the values: CI masks secrets, but a
+// stray echo would still leak them into local build logs.
+function stageIntegrationKeys() {
+  const staged = stageBundledIntegrationKeys(path.join(repoRoot, 'dist'));
+  process.stdout.write(
+    staged.length > 0
+      ? `Staged bundled integration keys: ${staged.join(', ')}\n`
+      : `No bundled integration keys (${TMDB_API_KEY_ENV} unset)\n`,
+  );
+}
+
 function main() {
  fs.cpSync(path.join(rendererSourceDir, 'fonts'), path.join(repoRoot, 'dist', 'fonts'), {
    recursive: true,
@@ -106,6 +118,7 @@ function main() {
  copySettingsAssets();
  copySyncUiAssets();
  buildMacosHelper();
+  stageIntegrationKeys();
 }

 main();
@@ -1,8 +1,10 @@
 import assert from 'node:assert/strict';
-import { resolve } from 'node:path';
+import { mkdirSync, mkdtempSync, rmSync, writeFileSync } from 'node:fs';
+import { tmpdir } from 'node:os';
+import { join, resolve } from 'node:path';
 import test from 'node:test';

-import { mergeLcovReports, resolveCoverageDir } from './run-coverage-lane';
+import { mergeLcovReports, resolveCoverageDir, runCoverageLane } from './run-coverage-lane';

 test('mergeLcovReports combines duplicate source-file counters across shard outputs', () => {
  const merged = mergeLcovReports([
@@ -72,3 +74,33 @@ test('resolveCoverageDir keeps coverage output inside the repository', () => {
  assert.throws(() => resolveCoverageDir(repoRoot, ['--coverage-dir', '../escape']));
  assert.throws(() => resolveCoverageDir(repoRoot, ['--coverage-dir', '/tmp/escape']));
 });
+
+test('runCoverageLane returns a failure when a discovered test fails', () => {
+  const repoRoot = mkdtempSync(join(tmpdir(), 'subminer-coverage-failure-'));
+  try {
+    mkdirSync(join(repoRoot, 'src'));
+    writeFileSync(
+      join(repoRoot, 'src', 'failure.test.ts'),
+      [
+        "import assert from 'node:assert/strict';",
+        "import test from 'node:test';",
+        '',
+        "test('intentional coverage failure', () => {",
+        "  assert.fail('coverage runner must propagate this failure');",
+        '});',
+        '',
+      ].join('\n'),
+    );
+
+    assert.notEqual(
+      runCoverageLane({
+        repoRootDir: repoRoot,
+        argv: ['bun-src-full', '--coverage-dir', 'coverage/test-src'],
+        stdio: 'pipe',
+      }),
+      0,
+    );
+  } finally {
+    rmSync(repoRoot, { recursive: true, force: true });
+  }
+});
@@ -201,14 +201,18 @@ export function mergeLcovReports(reports: string[]): string {
  return chunks.length > 0 ? `${chunks.join('\n')}\n` : '';
 }

-function runCoverageLane(): number {
-  const laneName = process.argv[2];
+export function runCoverageLane(
+  options: { repoRootDir?: string; argv?: string[]; stdio?: 'inherit' | 'pipe' } = {},
+): number {
+  const repoRootDir = options.repoRootDir ?? repoRoot;
+  const argv = options.argv ?? process.argv.slice(2);
+  const laneName = argv[0];
  if (laneName === undefined) {
    process.stderr.write('Missing coverage lane name\n');
    return 1;
  }

-  const coverageDir = resolveCoverageDir(repoRoot, process.argv.slice(3));
+  const coverageDir = resolveCoverageDir(repoRootDir, argv.slice(1));
  const shardRoot = join(coverageDir, '.shards');
  mkdirSync(coverageDir, { recursive: true });
  rmSync(shardRoot, { recursive: true, force: true });
@@ -216,7 +220,7 @@ function runCoverageLane(): number {

  let files: string[];
  try {
-    files = collectLaneFiles(repoRoot, laneName);
+    files = collectLaneFiles(repoRootDir, laneName);
  } catch (error) {
    process.stderr.write(`${error instanceof Error ? error.message : error}\n`);
    return 1;
@@ -230,8 +234,8 @@ function runCoverageLane(): number {
        'bun',
        ['test', '--coverage', '--coverage-reporter=lcov', '--coverage-dir', shardDir, `./${file}`],
        {
-          cwd: repoRoot,
-          stdio: 'inherit',
+          cwd: repoRootDir,
+          stdio: options.stdio ?? 'inherit',
        },
      );

@@ -253,7 +257,7 @@ function runCoverageLane(): number {

    writeFileSync(join(coverageDir, 'lcov.info'), mergeLcovReports(reports), 'utf8');
    process.stdout.write(
-      `Merged LCOV written to ${relative(repoRoot, join(coverageDir, 'lcov.info'))}\n`,
+      `Merged LCOV written to ${relative(repoRootDir, join(coverageDir, 'lcov.info'))}\n`,
    );
    return 0;
  } finally {
@@ -14,7 +14,12 @@ delete env.ELECTRON_RUN_AS_NODE;
 try {
  const result = spawnSync(
    require('electron'),
-    [fileURLToPath(new URL('./smoke-package.cjs', import.meta.url)), path.resolve(resources)],
+    [
+      fileURLToPath(new URL('./smoke-package.cjs', import.meta.url)),
+      path.resolve(resources),
+      // CI runners lack a setuid chrome-sandbox; this harness never loads remote content.
+      ...(process.platform === 'linux' ? ['--no-sandbox'] : []),
+    ],
    { env, stdio: 'inherit', timeout: 75_000 },
  );
  if (result.error) throw result.error;
@@ -1,6 +1,7 @@
 // Run with the pinned Electron runtime against a finished app's resources folder.
 const { app, BrowserWindow, session } = require('electron');
 const fs = require('node:fs');
+const http = require('node:http');
 const path = require('node:path');
 const { createRequire } = require('node:module');
 const assert = require('node:assert/strict');
@@ -21,6 +22,39 @@ const timeout = setTimeout(() => {
  app.exit(1);
 }, 60_000);

+const STATIC_TYPES = {
+  '.html': 'text/html',
+  '.js': 'text/javascript',
+  '.css': 'text/css',
+  '.png': 'image/png',
+  '.svg': 'image/svg+xml',
+  '.woff2': 'font/woff2',
+  '.ttf': 'font/ttf',
+  '.json': 'application/json',
+};
+
+// The stats dashboard is served by the stats HTTP server in the app, so load it
+// over loopback HTTP from the packaged stats/dist and treat missing static
+// assets as failures. API routes are not part of this smoke and may 404.
+function serveStatsDist(root, failedRequests) {
+  const server = http.createServer((req, res) => {
+    const pathname = new URL(req.url, 'http://127.0.0.1').pathname;
+    const relative = pathname === '/' ? 'index.html' : pathname.slice(1);
+    try {
+      const body = fs.readFileSync(path.join(root, relative));
+      res.writeHead(200, {
+        'Content-Type': STATIC_TYPES[path.extname(relative)] ?? 'application/octet-stream',
+      });
+      res.end(body);
+    } catch {
+      if (!pathname.startsWith('/api/')) failedRequests.push(`${req.url}: missing static asset`);
+      res.writeHead(404).end();
+    }
+  });
+  server.listen(0, '127.0.0.1');
+  return server;
+}
+
 async function smoke() {
  await app.whenReady();
  const packagedRequire = createRequire(path.join(archive, 'package.json'));
@@ -50,10 +84,17 @@ async function smoke() {
  );
  assert(extension.id, 'Yomitan extension failed to load');
  const failedRequests = [];
-  session.defaultSession.webRequest.onErrorOccurred({ urls: ['file://*/*'] }, (details) => {
-    if (details.error !== 'net::ERR_ABORTED')
-      failedRequests.push(`${details.url}: ${details.error}`);
-  });
+  session.defaultSession.webRequest.onErrorOccurred(
+    { urls: ['file://*/*', 'http://127.0.0.1/*'] },
+    (details) => {
+      // Chromium probes the cache before fetching @font-face fonts; an uncached
+      // font reports ERR_CACHE_MISS and is then fetched normally.
+      if (!['net::ERR_ABORTED', 'net::ERR_CACHE_MISS'].includes(details.error))
+        failedRequests.push(`${details.url}: ${details.error}`);
+    },
+  );
+  const statsServer = serveStatsDist(path.join(archive, 'stats', 'dist'), failedRequests);
+  await once(statsServer, 'listening');
  for (const ui of ['renderer', 'settings', 'syncui', 'stats']) {
    const win = new BrowserWindow({
      show: false,
@@ -63,10 +104,12 @@ async function smoke() {
      },
    });
    try {
-      await win.loadFile(
-        path.join(archive, ui === 'stats' ? 'stats/dist/index.html' : `dist/${ui}/index.html`),
-      );
-      if (ui !== 'stats') {
+      if (ui === 'stats') {
+        await win.loadURL(`http://127.0.0.1:${statsServer.address().port}/`);
+        // Let in-flight font requests settle before the window goes away.
+        await win.webContents.executeJavaScript('document.fonts.ready.then(() => true)');
+      } else {
+        await win.loadFile(path.join(archive, `dist/${ui}/index.html`));
        const loaded = await win.webContents.executeJavaScript(
          `document.fonts.load('400 16px "M PLUS 1"', '日本語').then(fonts => fonts.length > 0 && fonts.every(font => font.status === 'loaded'))`,
        );
@@ -76,6 +119,7 @@ async function smoke() {
      win.destroy();
    }
  }
+  statsServer.close();
  assert.deepEqual(failedRequests, [], 'Packaged UI resources failed to load');
  console.log(
    'Package smoke passed: SQLite, platform FFI, texthooker, Yomitan loading, UI pages, shared Japanese font.',
@@ -1527,3 +1527,43 @@ test('AnkiIntegration.formatMiscInfoPattern avoids leaking Jellyfin api_key quer
  assert.equal(result, '[SubMiner] [Jellyfin/direct] Bocchi the Rock! - S01E02 (00:07:06)');
  assert.equal(result.includes('api_key='), false);
 });
+
+test('Anki metadata rejects a credential-bearing media title before metadata arrives', () => {
+  const integration = new AnkiIntegration(
+    { metadata: { pattern: '[SubMiner] %f | %F (%t)' } } as never,
+    {} as never,
+    {
+      currentVideoPath: 'https://jellyfin.example/Videos/item/stream?api_key=test-secret',
+      currentMediaTitle: 'stream?static=true&api_key=test-secret',
+      currentTimePos: 426,
+      send: () => true,
+    } as never,
+  );
+  const privateApi = integration as unknown as {
+    formatMiscInfoPattern: (fallbackFilename: string, startTimeSeconds?: number) => string;
+  };
+  const result = privateApi.formatMiscInfoPattern('stream?api_key=test-secret', 426);
+  assert.equal(result, '[SubMiner] Unknown media | Unknown media (00:07:06)');
+});
+
+test('AnkiIntegration.formatMiscInfoPattern treats ApiKey stream paths like legacy api_key ones', () => {
+  const integration = new AnkiIntegration(
+    { metadata: { pattern: '[SubMiner] %f (%t)' } } as never,
+    {} as never,
+    {
+      currentSubText: '',
+      currentVideoPath: 'stream?static=true&ApiKey=secret-token&MediaSourceId=ms-1',
+      currentTimePos: 426,
+      currentSubStart: 426,
+      currentSubEnd: 428,
+      currentMediaTitle: '[Jellyfin/direct] Bocchi the Rock! - S01E02',
+      send: () => true,
+    } as unknown as never,
+  );
+  const privateApi = integration as unknown as {
+    formatMiscInfoPattern: (fallbackFilename: string, startTimeSeconds?: number) => string;
+  };
+  const result = privateApi.formatMiscInfoPattern('audio_123.mp3', 426);
+  assert.equal(result, '[SubMiner] [Jellyfin/direct] Bocchi the Rock! - S01E02 (00:07:06)');
+  assert.equal(result.includes('ApiKey='), false);
+});
@@ -32,6 +32,7 @@ import {
  type MediaTimingReviewRequest,
 } from './types/anki';
 import { AiConfig } from './types/integrations';
+import { sanitizeMediaTitle } from './shared/media-identity';
 import type { KnownWordMaturityTier } from './types/subtitle';
 import { MpvClient } from './types/runtime';
 import { OPEN_ANKI_CARD_ACTION_ID } from './types/notification';
@@ -184,7 +185,7 @@ function extractFilenameFromMediaPath(rawPath: string): string {
 function shouldPreferMediaTitleForMiscInfo(rawPath: string, filename: string): boolean {
  const loweredPath = rawPath.toLowerCase();
  const loweredFilename = filename.toLowerCase();
-  if (loweredPath.includes('api_key=')) {
+  if (loweredPath.includes('api_key=') || loweredPath.includes('apikey=')) {
    return true;
  }
  if (loweredPath.startsWith('http://') || loweredPath.startsWith('https://')) {
@@ -1117,9 +1118,11 @@ export class AnkiIntegration {
      return null;
    }
    const mediaRange = this.getSubtitleMediaRange(context);
-    const timestamp = context
-      ? mediaRange.startTime + (mediaRange.endTime - mediaRange.startTime) / 2
-      : this.mpvClient.currentTimePos || 0;
+    const timestamp =
+      context?.screenshotTime ??
+      (context
+        ? mediaRange.startTime + (mediaRange.endTime - mediaRange.startTime) / 2
+        : this.mpvClient.currentTimePos || 0);

    if (this.config.media?.imageType === 'avif') {
      return this.mediaGenerator.generateAnimatedImage(
@@ -1165,11 +1168,13 @@ export class AnkiIntegration {
    }

    const videoFilename = extractFilenameFromMediaPath(mediaPath);
-    const resolvedMediaTitle = trimToNonEmptyString(mediaTitle);
+    const resolvedMediaTitle = sanitizeMediaTitle(mediaTitle);
    const filenameWithExt =
      (shouldPreferMediaTitleForMiscInfo(mediaPath, videoFilename)
-        ? resolvedMediaTitle || videoFilename
-        : videoFilename || resolvedMediaTitle) || fallbackFilename;
+        ? resolvedMediaTitle || 'Unknown media'
+        : sanitizeMediaTitle(videoFilename) || resolvedMediaTitle) ||
+      sanitizeMediaTitle(fallbackFilename) ||
+      'Unknown media';
    const filenameWithoutExt = filenameWithExt.replace(/\.[^.]+$/, '');

    const currentTimePos =
@@ -1794,6 +1799,8 @@ export class AnkiIntegration {
      ...request,
      audioPadding: Math.max(0, this.config.media.audioPadding ?? 0),
      maxMediaDuration: Math.max(0, this.config.media.maxMediaDuration ?? 30),
+      screenshotEnabled:
+        this.config.media.generateImage !== false && this.config.media.imageType !== 'avif',
    });
  }

@@ -14,7 +14,8 @@ test('resolveAnimatedImageLeadInSeconds sums configured word audio durations for
  const leadInSeconds = await resolveAnimatedImageLeadInSeconds({
    config: {
      fields: {
-        audio: 'ExpressionAudio',
+        audio: 'SentenceAudio',
+        wordAudio: 'Pronunciation',
      },
      media: {
        imageType: 'avif',
@@ -25,7 +26,8 @@ test('resolveAnimatedImageLeadInSeconds sums configured word audio durations for
    noteInfo: {
      noteId: 42,
      fields: {
-        ExpressionAudio: {
+        SentenceAudio: { value: '[sound:sentence.mp3]' },
+        Pronunciation: {
          value: '[sound:word.mp3][sound:alt.ogg]',
        },
      },
@@ -121,3 +123,32 @@ test('resolveAnimatedImageLeadInSeconds falls back to zero when sync is disabled

  assert.equal(leadInSeconds, 0);
 });
+
+for (const sentenceAudio of ['', '[sound:sentence.mp3]']) {
+  test(`word audio defaults independently of sentence audio (${sentenceAudio ? 'existing' : 'new'} note)`, async () => {
+    const retrieved: string[] = [];
+    const leadInSeconds = await resolveAnimatedImageLeadInSeconds({
+      config: {
+        fields: { audio: 'SentenceAudio' },
+        media: { imageType: 'avif' },
+      },
+      noteInfo: {
+        noteId: 42,
+        fields: {
+          ExpressionAudio: { value: '[sound:word.mp3]' },
+          SentenceAudio: { value: sentenceAudio },
+        },
+      },
+      resolveConfiguredFieldName: (noteInfo, ...preferredNames) =>
+        preferredNames.find((name) => name !== undefined && name in noteInfo.fields) ?? null,
+      retrieveMediaFileBase64: async (filename) => {
+        retrieved.push(filename);
+        return 'd29yZA==';
+      },
+      probeAudioDurationSeconds: async (_buffer, filename) => (filename === 'word.mp3' ? 0.6 : 4),
+    });
+
+    assert.equal(leadInSeconds, 0.6);
+    assert.deepEqual(retrieved, ['word.mp3']);
+  });
+}
@@ -97,8 +97,8 @@ export async function resolveAnimatedImageLeadInSeconds<TNoteInfo extends NoteIn

  const wordAudioFieldName = resolveConfiguredFieldName(
    noteInfo,
-    config.fields?.audio,
-    DEFAULT_ANKI_CONNECT_CONFIG.fields.audio,
+    config.fields?.wordAudio,
+    DEFAULT_ANKI_CONNECT_CONFIG.fields.wordAudio,
  );
  if (!wordAudioFieldName) {
    return 0;
@@ -160,6 +160,52 @@ test('manual clipboard subtitle update replaces audio in the configured field',
  );
 });

+test('manual clipboard mining treats a zero media duration cap as unlimited', async () => {
+  const audioRanges: Array<{ start: number; end: number; padding: number | undefined }> = [];
+  const scenarios = [
+    { maxMediaDuration: 0, expectedEnd: 14 },
+    { maxMediaDuration: 1, expectedEnd: 13 },
+  ];
+
+  for (const scenario of scenarios) {
+    const { service } = createManualUpdateService({
+      getConfig: () =>
+        ({
+          deck: 'Mining',
+          fields: {
+            word: 'Expression',
+            sentence: 'Sentence',
+            audio: 'ExpressionAudio',
+          },
+          media: {
+            generateAudio: true,
+            generateImage: false,
+            audioPadding: 0.25,
+            maxMediaDuration: scenario.maxMediaDuration,
+          },
+          behavior: {},
+          ai: false,
+        }) as AnkiConnectConfig,
+      mediaGenerator: {
+        generateAudio: async (_path, start, end, padding) => {
+          audioRanges.push({ start, end, padding });
+          return Buffer.from('audio');
+        },
+        generateScreenshot: async () => null,
+        generateAnimatedImage: async () => null,
+      },
+    });
+
+    await service.updateLastAddedFromClipboard('字幕');
+
+    assert.deepEqual(audioRanges.at(-1), {
+      start: 12,
+      end: scenario.expectedEnd,
+      padding: 0.25,
+    });
+  }
+});
+
 test('manual clipboard word-card update uses configured fields with Lapis and Kiku enabled', async () => {
  const { service, updatedFields } = createManualUpdateService({
    getConfig: () =>
@@ -290,6 +290,9 @@ test('CardCreationService keeps updating after recordCardsMinedCallback throws',
 });

 test('CardCreationService uses stream-open-filename for remote media generation', async () => {
+  let reviewing = false;
+  const audioRanges: number[][] = [];
+  const imageTimes: number[] = [];
  const audioPaths: string[] = [];
  const imagePaths: string[] = [];
  const recordMediaPath = (mediaInput: MediaInput): string =>
@@ -319,6 +322,10 @@ test('CardCreationService uses stream-open-filename for remote media generation'
        behavior: {},
        ai: false,
      }) as AnkiConnectConfig,
+    reviewMediaTiming: async () =>
+      reviewing
+        ? { action: 'confirm', startTime: 0.2, endTime: 0.8, screenshotTime: 3.125 }
+        : { action: 'use-original' },
    getAiConfig: () => ({}),
    getTimingTracker: () => ({}) as never,
    getMpvClient: () =>
@@ -349,16 +356,18 @@ test('CardCreationService uses stream-open-filename for remote media generation'
      ],
      updateNoteFields: async () => undefined,
      storeMediaFile: async () => undefined,
-      findNotes: async () => [],
+      findNotes: async () => [42],
      retrieveMediaFile: async () => '',
      deleteNotes: async () => undefined,
    },
    mediaGenerator: {
-      generateAudio: async (path) => {
+      generateAudio: async (path, start, end, padding) => {
+        audioRanges.push([start, end, padding ?? -1]);
        audioPaths.push(recordMediaPath(path));
        return Buffer.from('audio');
      },
-      generateScreenshot: async (path) => {
+      generateScreenshot: async (path, timestamp) => {
+        imageTimes.push(timestamp);
        imagePaths.push(recordMediaPath(path));
        return Buffer.from('image');
      },
@@ -406,6 +415,14 @@ test('CardCreationService uses stream-open-filename for remote media generation'
  assert.equal(created, true);
  assert.deepEqual(audioPaths, [audioUrl]);
  assert.deepEqual(imagePaths, [videoUrl]);
+  reviewing = true;
+  assert.equal(await service.createSentenceCard('テスト', 0, 1), true);
+  assert.deepEqual(audioRanges.at(-1), [0.2, 0.8, 0]);
+  assert.equal(imageTimes.at(-1), 3.125);
+  await service.markLastCardAsAudioCard();
+  assert.equal(imageTimes.length, 3);
+  assert.deepEqual(audioRanges.at(-1), [0.2, 0.8, 0]);
+  assert.equal(imageTimes.at(-1), 3.125);
 });

 test('CardCreationService does not use mpv stream indexes for ready cached YouTube media', async () => {
@@ -21,6 +21,7 @@ import {
  resolveAudioStreamIndexForMediaGeneration,
  type MediaGenerationInputResolverOptions,
 } from './media-source';
+import { clampMediaEndTime } from './media-duration';
 import { resolveWordCardKind } from './note-field-utils';
 import type { PendingYoutubeMediaUpdate } from './pending-youtube-media';
 import { resolveMpvVolumeScale } from './mpv-volume';
@@ -233,11 +234,12 @@ export class CardCreationService {
      let rangeEnd = Math.max(...timings.map((entry) => entry.endTime));

      const maxMediaDuration = this.deps.getConfig().media?.maxMediaDuration ?? 30;
-      if (maxMediaDuration > 0 && rangeEnd - rangeStart > maxMediaDuration) {
+      const cappedRangeEnd = clampMediaEndTime(rangeStart, rangeEnd, maxMediaDuration);
+      if (cappedRangeEnd !== rangeEnd) {
        log.warn(
          `Media range ${(rangeEnd - rangeStart).toFixed(1)}s exceeds cap of ${maxMediaDuration}s, clamping`,
        );
-        rangeEnd = rangeStart + maxMediaDuration;
+        rangeEnd = cappedRangeEnd;
      }

      this.deps.showOsdNotification('Updating card from clipboard...');
@@ -437,9 +439,7 @@ export class CardCreationService {
      }

      const maxMediaDuration = this.deps.getConfig().media?.maxMediaDuration ?? 30;
-      if (maxMediaDuration > 0 && endTime - startTime > maxMediaDuration) {
-        endTime = startTime + maxMediaDuration;
-      }
+      endTime = clampMediaEndTime(startTime, endTime, maxMediaDuration);

      this.deps.showOsdNotification('Marking card as audio card...');
      await this.deps.withUpdateProgress('Marking audio card', async () => {
@@ -536,6 +536,7 @@ export class CardCreationService {
              endTime,
              animatedLeadInSeconds,
              exactReviewedRange,
+              timingDecision.action === 'confirm' ? timingDecision.screenshotTime : undefined,
            );

            const imageField = this.deps.getConfig().fields?.image;
@@ -599,11 +600,12 @@ export class CardCreationService {
    }

    const maxMediaDuration = this.deps.getConfig().media?.maxMediaDuration ?? 30;
-    if (maxMediaDuration > 0 && endTime - startTime > maxMediaDuration) {
+    const cappedEndTime = clampMediaEndTime(startTime, endTime, maxMediaDuration);
+    if (cappedEndTime !== endTime) {
      log.warn(
        `Sentence card media range ${(endTime - startTime).toFixed(1)}s exceeds cap of ${maxMediaDuration}s, clamping`,
      );
-      endTime = startTime + maxMediaDuration;
+      endTime = cappedEndTime;
    }

    try {
@@ -796,6 +798,9 @@ export class CardCreationService {
            generateImage,
            volumeScale,
            ...(exactReviewedRange ? { mediaPaddingSeconds: 0 } : {}),
+            ...(timingDecision.action === 'confirm' && timingDecision.screenshotTime !== undefined
+              ? { screenshotTime: timingDecision.screenshotTime }
+              : {}),
          });
          await this.deps.showNotification(noteId, label, 'media queued');
          return true;
@@ -840,6 +845,7 @@ export class CardCreationService {
              endTime,
              0,
              exactReviewedRange,
+              timingDecision.action === 'confirm' ? timingDecision.screenshotTime : undefined,
            );

            const imageField = config.fields?.image;
@@ -922,15 +928,16 @@ export class CardCreationService {
    endTime: number,
    animatedLeadInSeconds = 0,
    exactReviewedRange = false,
+    screenshotTime?: number,
  ): Promise<Buffer | null> {
    const mpvClient = this.deps.getMpvClient();
    if (!mpvClient) {
      return null;
    }

-    const timestamp = exactReviewedRange
-      ? startTime + (endTime - startTime) / 2
-      : mpvClient.currentTimePos || 0;
+    const timestamp =
+      screenshotTime ??
+      (exactReviewedRange ? startTime + (endTime - startTime) / 2 : mpvClient.currentTimePos || 0);

    if (this.deps.getConfig().media?.imageType === 'avif') {
      let imageStart = startTime;
@@ -0,0 +1,10 @@
+/** Zero or a negative cap leaves the requested end time unchanged. */
+export function clampMediaEndTime(
+  startTime: number,
+  endTime: number,
+  maxMediaDuration: number,
+): number {
+  return maxMediaDuration > 0 && endTime - startTime > maxMediaDuration
+    ? startTime + maxMediaDuration
+    : endTime;
+}
@@ -729,6 +729,7 @@ test('NoteUpdateWorkflow uses the combined review sentence for the card and medi
    startTime: 2,
    endTime: 7,
    text: 'previous-line current-line next-line',
+    screenshotTime: 8,
  });
  harness.deps.generateAudio = async (context) => {
    audioContexts.push(context);
@@ -745,6 +746,7 @@ test('NoteUpdateWorkflow uses the combined review sentence for the card and medi
  assert.equal(audioContexts[0]?.startTime, 2);
  assert.equal(audioContexts[0]?.endTime, 7);
  assert.equal(audioContexts[0]?.mediaPaddingSeconds, 0);
+  assert.equal(audioContexts[0]?.screenshotTime, 8);
 });

 test('NoteUpdateWorkflow keeps cache unchanged and reports when deletion fails', async () => {
@@ -253,6 +253,9 @@ export class NoteUpdateWorkflow {
            startTime: timingDecision.startTime,
            endTime: timingDecision.endTime,
            mediaPaddingSeconds: 0,
+            ...(timingDecision.screenshotTime !== undefined
+              ? { screenshotTime: timingDecision.screenshotTime }
+              : {}),
          };
        } else if (timingDecision.action === 'skip-media') {
          skipMedia = true;
@@ -51,6 +51,41 @@ function createDeps(
  return deps;
 }

+test('queued media keeps a chosen screenshot separate from the reviewed audio range', async () => {
+  const screenshots: number[] = [];
+  const audioRanges: number[][] = [];
+  const deps = createDeps();
+  deps.client.notesInfo = async () => [{ noteId: 42, fields: { Picture: { value: '' } } }];
+  deps.mediaGenerator.generateScreenshot = async (_media, time) => {
+    screenshots.push(time);
+    return Buffer.from('image');
+  };
+  deps.mediaGenerator.generateAudio = async (_media, start, end, padding) => {
+    audioRanges.push([start, end, padding ?? -1]);
+    return Buffer.from('audio');
+  };
+  const queue = new PendingYoutubeMediaQueue(deps);
+  assert.equal(
+    await queue.queueFromNote({
+      noteId: 42,
+      noteInfo: { noteId: 42, fields: {} },
+      label: 'test',
+      context: {
+        source: 'overlay',
+        text: '字幕',
+        startTime: 1,
+        endTime: 2,
+        mediaPaddingSeconds: 0,
+        screenshotTime: 3.125,
+      },
+    }),
+    true,
+  );
+  await queue.handleReady('https://youtu.be/abc123', '/cache/video.mkv');
+  assert.deepEqual(screenshots, [3.125]);
+  assert.deepEqual(audioRanges, [[1, 2, 0]]);
+});
+
 test('PendingYoutubeMediaQueue treats cache lookup failures as an immediate generation fallback', async () => {
  const deps = createDeps({
    getCachedMediaPath: async () => {
@@ -147,6 +147,9 @@ export class PendingYoutubeMediaQueue {
      generateAudio: shouldGenerateAudio(config),
      generateImage: shouldGenerateImage(config),
      volumeScale,
+      ...(job.context?.screenshotTime !== undefined
+        ? { screenshotTime: job.context.screenshotTime }
+        : {}),
      ...(job.context?.mediaPaddingSeconds !== undefined
        ? { mediaPaddingSeconds: job.context.mediaPaddingSeconds }
        : {}),
@@ -320,6 +323,7 @@ export class PendingYoutubeMediaQueue {
          job.endTime,
          animatedLeadInSeconds,
          job.mediaPaddingSeconds,
+          job.screenshotTime,
        );
        if (imageBuffer) {
          await this.deps.client.storeMediaFile(imageFilename, imageBuffer);
@@ -381,6 +385,7 @@ export class PendingYoutubeMediaQueue {
    endTime: number,
    animatedLeadInSeconds = 0,
    mediaPaddingSeconds?: number,
+    screenshotTime?: number,
  ): Promise<Buffer | null> {
    const config = this.deps.getConfig();
    if (config.media?.imageType === 'avif') {
@@ -399,7 +404,7 @@ export class PendingYoutubeMediaQueue {
      );
    }

-    const timestamp = startTime + (endTime - startTime) / 2;
+    const timestamp = screenshotTime ?? startTime + (endTime - startTime) / 2;
    return this.deps.mediaGenerator.generateScreenshot(videoPath, timestamp, {
      format: config.media?.imageFormat as 'jpg' | 'png' | 'webp',
      quality: config.media?.imageQuality,
@@ -11,6 +11,7 @@ export interface PendingYoutubeMediaUpdate {
  generateImage: boolean;
  volumeScale?: number;
  mediaPaddingSeconds?: number;
+  screenshotTime?: number;
 }

 function trimToNonEmptyString(value: unknown): string | null {
@@ -19,6 +19,23 @@ test('package scripts expose a sharded maintained source coverage lane with lcov
  );
 });

+test('source and coverage scripts discover the same maintained source lane', () => {
+  const sourceLane = packageJson.scripts['test:src']?.match(/run-test-lane\.mjs\s+([^\s]+)/)?.[1];
+  const coverageLane = packageJson.scripts['test:coverage:src']?.match(
+    /run-coverage-lane\.ts\s+([^\s]+)/,
+  )?.[1];
+
+  assert.equal(sourceLane, 'bun-src-full');
+  assert.equal(coverageLane, sourceLane);
+});
+
+test('environment suite owns launcher smoke execution', () => {
+  assert.match(
+    packageJson.scripts['test:env'] ?? '',
+    /^bun run test:launcher:smoke:src && bun run test:plugin:src && bun run test:immersion:sqlite:src$/,
+  );
+});
+
 test('ci delegates its gate instead of duplicating quality steps', () => {
  assert.match(
    ciWorkflow,
@@ -2818,6 +2818,29 @@ test('forces Senren off when Kiku is also enabled and validates Senren fieldGrou
  );
 });

+test('warns and falls back when isSenren.enabled is not boolean', () => {
+  const dir = makeTempDir();
+  fs.writeFileSync(
+    path.join(dir, 'config.jsonc'),
+    `{
+      "ankiConnect": {
+        "isSenren": { "enabled": "true" }
+      }
+    }`,
+    'utf-8',
+  );
+
+  const service = new ConfigService(dir);
+
+  assert.equal(
+    service.getConfig().ankiConnect.isSenren.enabled,
+    DEFAULT_CONFIG.ankiConnect.isSenren.enabled,
+  );
+  assert.ok(
+    service.getWarnings().some((warning) => warning.path === 'ankiConnect.isSenren.enabled'),
+  );
+});
+
 test('accepts valid ankiConnect knownWords deck object', () => {
  const dir = makeTempDir();
  fs.writeFileSync(
@@ -42,6 +42,7 @@ const {
  ankiConnect,
  jimaku,
  tsukihime,
+  tmdb,
  anilist,
  mpv,
  yomitan,
@@ -76,6 +77,7 @@ export const DEFAULT_CONFIG: ResolvedConfig = {
  auto_start_overlay,
  jimaku,
  tsukihime,
+  tmdb,
  anilist,
  mpv,
  yomitan,
@@ -6,6 +6,7 @@ export const INTEGRATIONS_DEFAULT_CONFIG: Pick<
  | 'ankiConnect'
  | 'jimaku'
  | 'tsukihime'
+  | 'tmdb'
  | 'anilist'
  | 'mpv'
  | 'yomitan'
@@ -29,6 +30,7 @@ export const INTEGRATIONS_DEFAULT_CONFIG: Pick<
    fields: {
      word: 'Expression',
      audio: 'ExpressionAudio',
+      wordAudio: 'ExpressionAudio',
      image: 'Picture',
      sentence: 'Sentence',
      miscInfo: 'MiscInfo',
@@ -112,6 +114,10 @@ export const INTEGRATIONS_DEFAULT_CONFIG: Pick<
    apiBaseUrl: 'https://api.tsukihime.org/v1',
    maxSearchResults: 10,
  },
+  tmdb: {
+    apiKey: '',
+    apiKeyCommand: '',
+  },
  mpv: {
    executablePath: '',
    launchMode: 'normal',
@@ -82,6 +82,13 @@ export function buildIntegrationConfigOptionRegistry(
      defaultValue: defaultConfig.ankiConnect.fields.audio,
      description: 'Card field that receives generated sentence audio.',
    },
+    {
+      path: 'ankiConnect.fields.wordAudio',
+      kind: 'string',
+      defaultValue: defaultConfig.ankiConnect.fields.wordAudio,
+      description:
+        'Existing word-audio field read to time the frozen first frame of animated images. This mapping is only used for synchronization.',
+    },
    {
      path: 'ankiConnect.fields.image',
      kind: 'string',
@@ -288,7 +295,7 @@ export function buildIntegrationConfigOptionRegistry(
      path: 'ankiConnect.media.maxMediaDuration',
      kind: 'number',
      defaultValue: defaultConfig.ankiConnect.media.maxMediaDuration,
-      description: 'Maximum allowed media clip duration in seconds.',
+      description: 'Maximum allowed media clip duration in seconds. 0 disables the cap.',
    },
    {
      path: 'ankiConnect.knownWords.matchMode',
@@ -472,6 +479,20 @@ export function buildIntegrationConfigOptionRegistry(
      defaultValue: defaultConfig.tsukihime.maxSearchResults,
      description: 'Maximum TsukiHime search results returned.',
    },
+    {
+      path: 'tmdb.apiKey',
+      kind: 'string',
+      defaultValue: defaultConfig.tmdb.apiKey,
+      description:
+        'Your own TMDB API key or read access token for live-action posters and synopses in the stats Library. Release builds bundle a project key, so set this only to use your own quota or when running from source (free under Settings > API on themoviedb.org).',
+    },
+    {
+      path: 'tmdb.apiKeyCommand',
+      kind: 'string',
+      defaultValue: defaultConfig.tmdb.apiKeyCommand,
+      description:
+        'Shell command that prints the TMDB API key to stdout. Used instead of apiKey to avoid storing the key in plain text.',
+    },
    {
      path: 'anilist.enabled',
      kind: 'boolean',
@@ -144,7 +144,7 @@ const INTEGRATION_TEMPLATE_SECTIONS: ConfigTemplateSection[] = [
    title: 'AnkiConnect Integration',
    description: ['Automatic Anki updates and media generation options.'],
    notes: [
-      'Hot-reload: ankiConnect.ai.enabled, media.normalizeAudio/mirrorMpvVolume/reviewTiming, knownWords, nPlusOne, fields.word/audio/image/sentence/miscInfo, behavior.autoUpdateNewCards, isLapis.sentenceCardModel, isKiku.fieldGrouping, isSenren.fieldGrouping, and lapisKiku.wordCardKind update live while SubMiner is running.',
+      'Hot-reload: ankiConnect.ai.enabled, media.normalizeAudio/mirrorMpvVolume/reviewTiming, knownWords, nPlusOne, fields.word/audio/wordAudio/image/sentence/miscInfo, behavior.autoUpdateNewCards, isLapis.sentenceCardModel, isKiku.fieldGrouping, isSenren.fieldGrouping, and lapisKiku.wordCardKind update live while SubMiner is running.',
      'Shared AI provider transport settings are read from top-level ai and typically require restart.',
      'Most other AnkiConnect settings still require restart.',
    ],
@@ -164,6 +164,14 @@ const INTEGRATION_TEMPLATE_SECTIONS: ConfigTemplateSection[] = [
    notes: ['Hot-reload: TsukiHime changes apply to the next TsukiHime request.'],
    key: 'tsukihime',
  },
+  {
+    title: 'TMDB',
+    description: [
+      'TMDB (The Movie Database) metadata for live-action dramas and movies in the stats Library: posters, synopses, and grouping by show.',
+    ],
+    notes: ['Hot-reload: TMDB changes apply to the next TMDB request.'],
+    key: 'tmdb',
+  },
  {
    title: 'YouTube Playback Settings',
    description: [
@@ -0,0 +1,65 @@
+function pathStartsWith(path: string, prefix: string): boolean {
+  return path === prefix || path.startsWith(`${prefix}.`);
+}
+
+const HOT_RELOAD_ROOTS = ['subtitleStyle', 'keybindings', 'shortcuts', 'subtitleSidebar'] as const;
+
+const HOT_RELOAD_EXACT_OR_PREFIX_PATHS = [
+  'secondarySub.defaultMode',
+  'mpv.aniskipEnabled',
+  'mpv.aniskipButtonKey',
+  'ankiConnect.ai.enabled',
+  'stats.toggleKey',
+  'stats.markWatchedKey',
+  'logging.level',
+  'logging.rotation',
+  'logging.files',
+  'youtube.primarySubLanguages',
+  'ankiConnect.deck',
+  'ankiConnect.media.normalizeAudio',
+  'ankiConnect.media.mirrorMpvVolume',
+  'ankiConnect.media.reviewTiming',
+  'ankiConnect.behavior.autoUpdateNewCards',
+  'ankiConnect.knownWords.highlightEnabled',
+  'ankiConnect.knownWords.refreshMinutes',
+  'ankiConnect.knownWords.addMinedWordsImmediately',
+  'ankiConnect.knownWords.matchMode',
+  'ankiConnect.knownWords.decks',
+  'ankiConnect.nPlusOne.enabled',
+  'ankiConnect.nPlusOne.minSentenceWords',
+  'ankiConnect.fields.word',
+  'ankiConnect.fields.audio',
+  'ankiConnect.fields.wordAudio',
+  'ankiConnect.fields.image',
+  'ankiConnect.fields.sentence',
+  'ankiConnect.fields.miscInfo',
+  'ankiConnect.isLapis.sentenceCardModel',
+  'ankiConnect.isKiku.fieldGrouping',
+  'ankiConnect.isSenren.fieldGrouping',
+  'ankiConnect.lapisKiku.wordCardKind',
+] as const;
+
+export function getConfigHotReloadField(path: string): string | null {
+  for (const root of HOT_RELOAD_ROOTS) {
+    if (pathStartsWith(path, root)) {
+      return root;
+    }
+  }
+
+  for (const hotPath of HOT_RELOAD_EXACT_OR_PREFIX_PATHS) {
+    if (pathStartsWith(path, hotPath)) {
+      return hotPath;
+    }
+  }
+
+  // These consumers read the current config when the next operation starts.
+  if (
+    ['jimaku', 'tmdb', 'subsync', 'notifications', 'subtitleGeneration'].some((root) =>
+      pathStartsWith(path, root),
+    )
+  ) {
+    return path;
+  }
+
+  return null;
+}
@@ -32,6 +32,101 @@ test('media timing review is disabled by default and accepts a boolean override'
  assert.deepEqual(enabledContext.warnings, []);
 });

+test('invalid direct and field-grouping Anki values warn and keep defaults', () => {
+  const { context, warnings } = makeContext({
+    enabled: 'true',
+    url: 8765,
+    pollingRate: '3000',
+    deck: ['Mining'],
+    isKiku: {
+      enabled: 1,
+      fieldGrouping: 'sometimes',
+      deleteDuplicateInAuto: 'false',
+    },
+    isSenren: {
+      enabled: 'true',
+      fieldGrouping: false,
+      deleteDuplicateInAuto: 0,
+    },
+  });
+
+  applyAnkiConnectResolution(context);
+
+  assert.equal(context.resolved.ankiConnect.enabled, DEFAULT_CONFIG.ankiConnect.enabled);
+  assert.equal(context.resolved.ankiConnect.url, DEFAULT_CONFIG.ankiConnect.url);
+  assert.equal(context.resolved.ankiConnect.pollingRate, DEFAULT_CONFIG.ankiConnect.pollingRate);
+  assert.equal(context.resolved.ankiConnect.deck, DEFAULT_CONFIG.ankiConnect.deck);
+  assert.deepEqual(context.resolved.ankiConnect.isKiku, DEFAULT_CONFIG.ankiConnect.isKiku);
+  assert.deepEqual(context.resolved.ankiConnect.isSenren, DEFAULT_CONFIG.ankiConnect.isSenren);
+  assert.deepEqual(
+    warnings.map((warning) => warning.path),
+    [
+      'ankiConnect.enabled',
+      'ankiConnect.url',
+      'ankiConnect.pollingRate',
+      'ankiConnect.deck',
+      'ankiConnect.isKiku.enabled',
+      'ankiConnect.isKiku.deleteDuplicateInAuto',
+      'ankiConnect.isKiku.fieldGrouping',
+      'ankiConnect.isSenren.enabled',
+      'ankiConnect.isSenren.deleteDuplicateInAuto',
+      'ankiConnect.isSenren.fieldGrouping',
+    ],
+  );
+});
+
+test('accepts valid direct and field-grouping Anki values', () => {
+  const { context, warnings } = makeContext({
+    enabled: false,
+    url: 'http://127.0.0.1:9876',
+    pollingRate: 750,
+    deck: 'Mining',
+    isKiku: {
+      enabled: true,
+      fieldGrouping: 'manual',
+      deleteDuplicateInAuto: false,
+    },
+    isSenren: {
+      enabled: false,
+      fieldGrouping: 'disabled',
+      deleteDuplicateInAuto: false,
+    },
+  });
+
+  applyAnkiConnectResolution(context);
+
+  assert.equal(context.resolved.ankiConnect.enabled, false);
+  assert.equal(context.resolved.ankiConnect.url, 'http://127.0.0.1:9876');
+  assert.equal(context.resolved.ankiConnect.pollingRate, 750);
+  assert.equal(context.resolved.ankiConnect.deck, 'Mining');
+  assert.deepEqual(context.resolved.ankiConnect.isKiku, {
+    enabled: true,
+    fieldGrouping: 'manual',
+    deleteDuplicateInAuto: false,
+  });
+  assert.deepEqual(context.resolved.ankiConnect.isSenren, {
+    enabled: false,
+    fieldGrouping: 'disabled',
+    deleteDuplicateInAuto: false,
+  });
+  assert.deepEqual(warnings, []);
+});
+
+test('ignores unknown Anki keys without warning or admitting them to resolved config', () => {
+  const { context, warnings } = makeContext({
+    futureOption: { enabled: true },
+    isKiku: { futureGroupingOption: 'future' },
+    isSenren: { futureGroupingOption: 'future' },
+  });
+
+  applyAnkiConnectResolution(context);
+
+  assert.equal(Object.hasOwn(context.resolved.ankiConnect, 'futureOption'), false);
+  assert.equal(Object.hasOwn(context.resolved.ankiConnect.isKiku, 'futureGroupingOption'), false);
+  assert.equal(Object.hasOwn(context.resolved.ankiConnect.isSenren, 'futureGroupingOption'), false);
+  assert.deepEqual(warnings, []);
+});
+
 test('modern media duration accepts zero as the disabled cap sentinel', () => {
  const disabledCap = makeContext({ media: { maxMediaDuration: 0 } });
  applyAnkiConnectResolution(disabledCap.context);
@@ -290,6 +385,28 @@ test('accepts ankiConnect.media.syncAnimatedImageToWordAudio override', () => {
  );
 });

+test('word audio mapping defaults and validates independently of sentence audio', () => {
+  for (const wordAudio of [undefined, 'Pronunciation', 7]) {
+    const { context, warnings } = makeContext({
+      fields: {
+        audio: 'SentenceAudio',
+        ...(wordAudio !== undefined ? { wordAudio } : {}),
+      },
+    });
+    applyAnkiConnectResolution(context);
+
+    assert.equal(context.resolved.ankiConnect.fields.audio, 'SentenceAudio');
+    assert.equal(
+      context.resolved.ankiConnect.fields.wordAudio,
+      typeof wordAudio === 'string' ? wordAudio : DEFAULT_CONFIG.ankiConnect.fields.wordAudio,
+    );
+    assert.deepEqual(
+      warnings.map((warning) => warning.path),
+      typeof wordAudio === 'number' ? ['ankiConnect.fields.wordAudio'] : [],
+    );
+  }
+});
+
 test('invalid modern Anki subtrees warn and keep resolved defaults', () => {
  const { context, warnings } = makeContext({
    fields: { word: 7 },
@@ -19,11 +19,11 @@ export function applyAnkiConnectResolution(context: ResolveContext): void {
  const media = isObject(ankiConnect.media) ? ankiConnect.media : {};
  const metadata = isObject(ankiConnect.metadata) ? ankiConnect.metadata : {};

-  initializeAnkiConnectResolution(context, ankiConnect);
+  initializeAnkiConnectResolution(context);
  applyAnkiModernResolution(context, ankiConnect, behavior, media);
  applyAnkiLegacyResolution(context, ankiConnect, behavior, fields, media, metadata);
  applyAnkiKnownWordsResolution(context, ankiConnect, behavior);
-  applyAnkiKikuResolution(context);
-  applyAnkiSenrenResolution(context);
+  applyAnkiKikuResolution(context, ankiConnect);
+  applyAnkiSenrenResolution(context, ankiConnect);
  applyAnkiLapisKikuResolution(context, ankiConnect);
 }
@@ -0,0 +1,58 @@
+import { DEFAULT_CONFIG } from '../../definitions';
+import type { ResolveContext } from '../context';
+import { asBoolean, asString } from '../shared';
+import { applyModernValue, asPositiveNumber } from './modern-value';
+
+export function applyAnkiBaseResolution(
+  context: ResolveContext,
+  ankiConnect: Record<string, unknown>,
+): void {
+  applyModernValue(
+    context,
+    ankiConnect,
+    'enabled',
+    'ankiConnect.enabled',
+    asBoolean,
+    DEFAULT_CONFIG.ankiConnect.enabled,
+    (value) => {
+      context.resolved.ankiConnect.enabled = value;
+    },
+    'Expected boolean.',
+  );
+  applyModernValue(
+    context,
+    ankiConnect,
+    'url',
+    'ankiConnect.url',
+    asString,
+    DEFAULT_CONFIG.ankiConnect.url,
+    (value) => {
+      context.resolved.ankiConnect.url = value;
+    },
+    'Expected string.',
+  );
+  applyModernValue(
+    context,
+    ankiConnect,
+    'pollingRate',
+    'ankiConnect.pollingRate',
+    asPositiveNumber,
+    DEFAULT_CONFIG.ankiConnect.pollingRate,
+    (value) => {
+      context.resolved.ankiConnect.pollingRate = value;
+    },
+    'Expected positive number.',
+  );
+  applyModernValue(
+    context,
+    ankiConnect,
+    'deck',
+    'ankiConnect.deck',
+    asString,
+    DEFAULT_CONFIG.ankiConnect.deck,
+    (value) => {
+      context.resolved.ankiConnect.deck = value;
+    },
+    'Expected string.',
+  );
+}
@@ -0,0 +1,53 @@
+import { DEFAULT_CONFIG } from '../../definitions';
+import type { ResolveContext } from '../context';
+import { asBoolean, isObject } from '../shared';
+import { applyModernValue } from './modern-value';
+
+type FieldGroupingConfigKey = 'isKiku' | 'isSenren';
+
+export function applyFieldGroupingConfigResolution(
+  context: ResolveContext,
+  ankiConnect: Record<string, unknown>,
+  key: FieldGroupingConfigKey,
+): void {
+  const source = ankiConnect[key];
+  if (!isObject(source)) {
+    if (source !== undefined) {
+      context.warn(
+        `ankiConnect.${key}`,
+        source,
+        DEFAULT_CONFIG.ankiConnect[key],
+        'Expected object.',
+      );
+    }
+    return;
+  }
+
+  for (const booleanKey of ['enabled', 'deleteDuplicateInAuto'] as const) {
+    applyModernValue(
+      context,
+      source,
+      booleanKey,
+      `ankiConnect.${key}.${booleanKey}`,
+      asBoolean,
+      DEFAULT_CONFIG.ankiConnect[key][booleanKey],
+      (value) => {
+        context.resolved.ankiConnect[key][booleanKey] = value;
+      },
+      'Expected boolean.',
+    );
+  }
+
+  applyModernValue(
+    context,
+    source,
+    'fieldGrouping',
+    `ankiConnect.${key}.fieldGrouping`,
+    (value) => (value === 'auto' || value === 'manual' || value === 'disabled' ? value : undefined),
+    DEFAULT_CONFIG.ankiConnect[key].fieldGrouping,
+    (value) => {
+      context.resolved.ankiConnect[key].fieldGrouping = value;
+    },
+    'Expected auto, manual, or disabled.',
+  );
+}
@@ -1,55 +1,8 @@
 import type { ResolveContext } from '../context';
-import { isObject } from '../shared';
-
-const LEGACY_KEYS = new Set([
-  'wordField',
-  'audioField',
-  'imageField',
-  'sentenceField',
-  'miscInfoField',
-  'miscInfoPattern',
-  'generateAudio',
-  'generateImage',
-  'imageType',
-  'imageFormat',
-  'imageQuality',
-  'imageMaxWidth',
-  'imageMaxHeight',
-  'animatedFps',
-  'animatedMaxWidth',
-  'animatedMaxHeight',
-  'animatedCrf',
-  'syncAnimatedImageToWordAudio',
-  'audioPadding',
-  'fallbackDuration',
-  'maxMediaDuration',
-  'overwriteAudio',
-  'overwriteImage',
-  'mediaInsertMode',
-  'highlightWord',
-  'notificationType',
-  'autoUpdateNewCards',
-]);
-
-export function initializeAnkiConnectResolution(
-  context: ResolveContext,
-  ankiConnect: Record<string, unknown>,
-): void {
-  const {
-    knownWords: _knownWordsConfigFromAnkiConnect,
-    nPlusOne: _nPlusOneConfigFromAnkiConnect,
-    ai: _ankiAiConfig,
-    ...ankiConnectWithoutKnownWordsOrNPlusOne
-  } = ankiConnect;
-  const ankiConnectWithoutLegacy = Object.fromEntries(
-    Object.entries(ankiConnectWithoutKnownWordsOrNPlusOne).filter(([key]) => !LEGACY_KEYS.has(key)),
-  );

+export function initializeAnkiConnectResolution(context: ResolveContext): void {
  context.resolved.ankiConnect = {
    ...context.resolved.ankiConnect,
-    ...(isObject(ankiConnectWithoutLegacy)
-      ? (ankiConnectWithoutLegacy as Partial<(typeof context.resolved)['ankiConnect']>)
-      : {}),
    fields: {
      ...context.resolved.ankiConnect.fields,
    },
@@ -73,15 +26,9 @@ export function initializeAnkiConnectResolution(
    },
    isKiku: {
      ...context.resolved.ankiConnect.isKiku,
-      ...(isObject(ankiConnect.isKiku)
-        ? (ankiConnect.isKiku as (typeof context.resolved)['ankiConnect']['isKiku'])
-        : {}),
    },
    isSenren: {
      ...context.resolved.ankiConnect.isSenren,
-      ...(isObject(ankiConnect.isSenren)
-        ? (ankiConnect.isSenren as (typeof context.resolved)['ankiConnect']['isSenren'])
-        : {}),
    },
    lapisKiku: {
      ...context.resolved.ankiConnect.lapisKiku,
@@ -1,19 +1,9 @@
-import { DEFAULT_CONFIG } from '../../definitions';
 import type { ResolveContext } from '../context';
+import { applyFieldGroupingConfigResolution } from './field-grouping-config';

-export function applyAnkiKikuResolution(context: ResolveContext): void {
-  if (
-    context.resolved.ankiConnect.isKiku.fieldGrouping !== 'auto' &&
-    context.resolved.ankiConnect.isKiku.fieldGrouping !== 'manual' &&
-    context.resolved.ankiConnect.isKiku.fieldGrouping !== 'disabled'
-  ) {
-    context.warn(
-      'ankiConnect.isKiku.fieldGrouping',
-      context.resolved.ankiConnect.isKiku.fieldGrouping,
-      DEFAULT_CONFIG.ankiConnect.isKiku.fieldGrouping,
-      'Expected auto, manual, or disabled.',
-    );
-    context.resolved.ankiConnect.isKiku.fieldGrouping =
-      DEFAULT_CONFIG.ankiConnect.isKiku.fieldGrouping;
-  }
+export function applyAnkiKikuResolution(
+  context: ResolveContext,
+  ankiConnect: Record<string, unknown>,
+): void {
+  applyFieldGroupingConfigResolution(context, ankiConnect, 'isKiku');
 }
@@ -7,7 +7,15 @@ export function applyModernFieldsResolution(
  context: ResolveContext,
  fields: Record<string, unknown>,
 ): void {
-  for (const key of ['word', 'audio', 'image', 'sentence', 'miscInfo', 'translation'] as const) {
+  for (const key of [
+    'word',
+    'audio',
+    'wordAudio',
+    'image',
+    'sentence',
+    'miscInfo',
+    'translation',
+  ] as const) {
    applyModernValue(
      context,
      fields,
@@ -1,6 +1,7 @@
 import type { ResolveContext } from '../context';
 import { isObject } from '../shared';
 import { applyAiResolution } from './ai';
+import { applyAnkiBaseResolution } from './base';
 import { applyLapisResolution } from './lapis';
 import { applyModernBehaviorResolution } from './modern-behavior';
 import { applyModernFieldsResolution } from './modern-fields';
@@ -18,6 +19,7 @@ export function applyAnkiModernResolution(
  const fields = isObject(ankiConnect.fields) ? ankiConnect.fields : {};
  const metadata = isObject(ankiConnect.metadata) ? ankiConnect.metadata : {};

+  applyAnkiBaseResolution(context, ankiConnect);
  applyModernFieldsResolution(context, fields);
  applyModernMediaResolution(context, media);
  applyModernBehaviorResolution(context, behavior);
@@ -1,21 +1,11 @@
-import { DEFAULT_CONFIG } from '../../definitions';
 import type { ResolveContext } from '../context';
+import { applyFieldGroupingConfigResolution } from './field-grouping-config';

-export function applyAnkiSenrenResolution(context: ResolveContext): void {
-  if (
-    context.resolved.ankiConnect.isSenren.fieldGrouping !== 'auto' &&
-    context.resolved.ankiConnect.isSenren.fieldGrouping !== 'manual' &&
-    context.resolved.ankiConnect.isSenren.fieldGrouping !== 'disabled'
-  ) {
-    context.warn(
-      'ankiConnect.isSenren.fieldGrouping',
-      context.resolved.ankiConnect.isSenren.fieldGrouping,
-      DEFAULT_CONFIG.ankiConnect.isSenren.fieldGrouping,
-      'Expected auto, manual, or disabled.',
-    );
-    context.resolved.ankiConnect.isSenren.fieldGrouping =
-      DEFAULT_CONFIG.ankiConnect.isSenren.fieldGrouping;
-  }
+export function applyAnkiSenrenResolution(
+  context: ResolveContext,
+  ankiConnect: Record<string, unknown>,
+): void {
+  applyFieldGroupingConfigResolution(context, ankiConnect, 'isSenren');

  // Kiku and Senren field grouping write incompatible markup into the same note
  // fields, so only one may be active; Kiku wins to preserve pre-existing setups.
Author	SHA1	Message	Date
sudacode	6a1d1211ae	fix(release): resolve style.css asar lookup with native separators The shared-font check still handed asar a forward-slash path, which fails on Windows the same way the inventory listing did.	2026-09-22 02:29:28 -07:00
sudacode	7585d430ae	fix(release): unbreak packaging workflow on all three platforms - Bump electron-builder to 26.16.1 so macOS signing passes the temp keychain password to set-key-partition-list (upstream #10066), which the macOS 26 runner image now enforces - Stat asar entries with their native path in the package audit; splitting on path.sep made every nested file look missing on Windows - Run the package smoke with --no-sandbox on Linux runners without a setuid chrome-sandbox - Load the packaged stats dashboard over loopback HTTP in the smoke, matching how the app serves it since #263, and ignore Chromium's cache-only font probe	2026-09-22 02:15:49 -07:00
sudacode	122ff9a3c5	chore(release): prepare v0.20.0-beta.1 Trim pending change fragments, drop the breaking flag from the stats request-safety change, and generate prerelease notes.	2026-09-22 01:29:58 -07:00
sudacode	786bd5af81	chore(yomitan): update vendored fork to upstream 26.9.8	2026-09-21 20:53:26 -07:00
sudacode	42d1384c68	fix(jellyfin): support modern authentication (#264 )	2026-09-21 20:40:57 -07:00
sudacode	1508863dbb	feat(stats): add TMDB metadata for live-action dramas in the Library (#252 )	2026-09-21 00:10:00 -07:00
sudacode	4dd30f44d1	chore(yomitan): update fork to upstream 26.9.8	2026-09-21 00:09:48 -07:00
sudacode	ab48a5678e	fix(stats): restrict local requests and serve the dashboard over HTTP (#263 )	2026-09-20 23:36:55 -07:00
sudacode	2f21582666	fix(overlay): cancel pending window transitions and timing reviews (#262 )	2026-09-20 23:35:48 -07:00
sudacode	383aed8bad	fix(stats): harden server lifecycle and verify compiled runtime (#261 )	2026-09-20 23:19:03 -07:00
sudacode	f9c4e892dc	fix(stats): reject malformed resource IDs before mutations (#259 )	2026-09-20 22:49:27 -07:00
sudacode	7959e1a4e5	fix(anki): honor unlimited duration in stats mining (#258 )	2026-09-20 22:49:12 -07:00
sudacode	917d52dd94	fix(config): validate AnkiConnect and field grouping settings (#257 )	2026-09-20 22:48:54 -07:00
sudacode	05425cc5db	fix(anki): separate word audio mapping for animation sync (#256 )	2026-09-20 22:07:30 -07:00
sudacode	8097ed27e5	feat(stats): separate YouTube channels in the Library (#253 )	2026-09-20 19:57:32 -07:00
sudacode	76e43d7e6d	fix(config): align live save feedback with hot reload policy (#255 )	2026-09-20 19:51:45 -07:00
Abdulrazzaq Alhendi	4f0762e840	feat(mining): add a screenshot frame picker to media review (#254 )	2026-09-20 19:39:30 -07:00
sudacode	026d495fac	feat(jimaku): add live-action subtitle search (#251 )	2026-09-18 22:50:13 -07:00
sudacode	dd76782d30	fix(subtitles): widen subtitle generation modal - Reduce text wrapping and scrolling while retaining small-window support - Document the responsive layout change	2026-09-18 21:05:52 -07:00
sudacode	a8c16147ad	feat(subtitles): use loaded subtitles to guide generation timing (#249 )	2026-09-18 00:24:42 -07:00
sudacode	186d4a0640	fix(jellyfin): fix jellyfin media metadata (#250 )	2026-09-18 00:24:25 -07:00