Track and run run-once SQL data migrations: entrypoint now checks _data_migrations before executing each prisma data-migration file, records successful runs, and skips already-applied scripts. Adds a Prisma DataMigration model mapped to _data_migrations and a new reset-works-table.sql migration to clear work tables for a dedup rebuild. Also improves dedup logic: extractSubtitle and subtitle-compatibility checks are added so series entries like "Series: Book A" vs "Series: Book B" are not collapsed, with accompanying unit tests for extraction and behavior.
Add an entrypoint step to execute idempotent SQL data migrations (prisma db execute) from prisma/data-migrations/*.sql so fixes that prisma db push doesn't handle are applied on startup. Add normalize-local-usernames.sql to normalize local users' plex_username and plex_id to lowercase. Update interactive search and search-indexers processor to prefer the user-provided/custom search title (searchTitle / effectiveSearchTitle) when ranking torrents and adjust debug logs to show the ranking title alongside the audiobook title/author for clearer diagnostics.
Replace default Anna's Archive base URL from https://annas-archive.li to https://annas-archive.gl across docs, UI components, API routes, processors, services, and tests. Add comprehensive tests for the admin manual-import API route and enhance the manual-import route to fetch missing ASIN details from Audnexus and create audiobook records with proper error handling and logging. Update related test expectations and FlareSolverr test usages to reflect the new default URL.
Introduce per-user configurable home page sections and a unified Audible cache/category model. Adds Prisma models (UserHomeSection, AudibleCacheCategory) and migrations to create tables and remove legacy popular/new_release flags; updates schema.prisma accordingly. Add API routes for user home sections, live Audible categories, and category-based audiobook listing, and refactor popular/new-releases/covers routes to read from AudibleCacheCategory. Frontend: new HomeSection component, HomeSectionConfigModal, useHomeSections hook, and homepage changes to render dynamic sections plus image fallback to a placeholder SVG. Also add placeholder_cover.svg and tests for home sections and the audible refresh processor.
When decrypting a user's API token fails, increment stats.errors and continue to the next shelf instead of proceeding. This ensures failed decryptions are tracked in metrics and prevents attempting to fetch data with an invalid token.
Drop the duplicated `id` alias from JWT payloads and related token generation across auth providers and endpoints. The TokenPayload interface no longer includes `id`; middleware now derives `user.id` from `sub` when attaching the authenticated user to requests. Update tests accordingly. This reduces redundancy and ensures the canonical user identifier is `sub`.
Introduce a shared useApiTokens hook to centralize API token CRUD and UI state (fetch, create, delete, copy, formatting). Refactor ApiTab and ApiTokensSection to consume the hook and remove duplicated logic. Add getInstanceUrl utility for client origin used in curl examples. Include an id alias in TokenPayload and add id into generated JWTs across auth routes and providers; update tests accordingly. Improve auth middleware typing and add debug logging around lastUsedAt updates. Add admin logging when creating a token with a role that differs from the target user's role.
Introduce full API token support: add a Prisma migration to create api_tokens table and indexes; add types, constants and a generateApiToken utility (hashed token + prefix). Update admin and user token routes to use the generator, enforce per-user active token caps, and integrate rate-limit checks. Add an interactive API docs page with TokenInput, EndpointCard and ResponseViewer components, plus a protected page route. Improve confirmation UX with an accessible ConfirmDialog (focus trap, Escape to close, animations) and wire confirm flows into admin/profile token sections; also update ConfirmModal to accept node messages. Add dialog CSS animations and enhance clipboard error handling. Update related middleware, utils and tests to reflect changes.
Normalize local account usernames by trimming and lowercasing across the stack. Added a Prisma migration to lowercase existing plex_username and rewrite local plex_id values for non-deleted accounts. Updated LocalAuthProvider, admin login route, and setup completion to use normalized usernames when looking up, creating, and storing users (including plexId `local-{username}`). Added/updated tests to assert case-insensitive lookups, storage of lowercased usernames/plexIds, and duplicate username rejection.
Query Audnexus (Audible) to backfill missing metadata during manual imports and file organization. Adds getAudibleService imports and calls to fetch audiobook details by ASIN, then backfills series, seriesPart, seriesAsin, year (from releaseDate) and narrator when missing and updates the DB. Failures are non-fatal and logged; logs were added to surface enrichment steps. Also uses the resolved series/seriesPart when building organization metadata.
Multiple fixes and improvements:
- src/app/api/user/hardcover-shelves/[id]/route.ts: Make token testing more robust by using the existing shelf.apiToken when no new token is provided, attempt decryption only when needed, and gracefully fall back on decryption errors.
- src/components/ui/AddShelfModal.tsx: Simplify token handling by passing the trimmed token directly to addHardcover (remove client-side 'Bearer ' stripping).
- src/components/ui/ManageShelfModal.tsx: Stabilize form reset effect by depending on shelf?.id to avoid unnecessary re-renders when the shelf object changes identity.
- src/components/ui/Modal.tsx: Simplify modal rendering by removing the mounted state and createPortal usage, cleaning up imports and rendering directly.
- src/lib/services/hardcover-api.service.ts: Add a logger, introduce a MAX_PAGES cap and page counters to prevent unbounded pagination loops, and log/break when the API returns errors during pagination.
These changes improve reliability (token handling and pagination safety), reduce unnecessary renders, and simplify modal lifecycle.
Replace the old hardcover sync usage with a new hardcover-api.service implementation that adds types, a reusable extractBooks helper, and paginated GraphQL queries (limit/offset) to fully fetch status and list books. Update API route import to use the new service. Fix ManageShelfModal to initialize rssUrl/listId as empty strings. Update tests to mock the new service and add encryption format helper mocking.
Introduce Hardcover provider support and consolidate per-provider book mapping tables into a unified BookMapping model. Adds two Prisma migrations (add_hardcover_shelves, unify_book_mappings), new backend services (hardcover-api, shelf-sync-core), and provider-specific sync logic and API routes for hardcover shelves with token/list validation. Frontend: new HardcoverForm component, refactor AddShelfModal to support Hardcover, hook updates, and small UI/accessibility tweaks. Also add documentation for Goodreads and Hardcover sync flows and update tests to cover scheduler/prisma helpers.
Introduce watched lists for series and authors end-to-end.
- Add DB migration to create watched_series and watched_authors tables with indexes and foreign keys.
- Implement API routes: GET/POST for listing/adding and DELETE by id for both /api/user/watched-series and /api/user/watched-authors. Validation, ownership checks, and immediate targeted job triggers are included.
- Add client hooks (useWatchedSeries, useWatchedAuthors) with add/delete helpers and SWR revalidation.
- Add UI components: WatchButton (toggle/confirm) and WatchedListsSection for profile display and removal UX.
- Add processor (check-watched-lists.processor) and service (watched-lists.service) to scrape Audible, deduplicate, check library ownership, and auto-create requests; supports targeted checks for newly watched items.
- Include tests for the watched-lists service.
These changes implement the watched-lists feature to let users watch series/authors and have the system automatically detect and request new releases.
- Add userId field to ApiToken schema (the user identity the token acts as)
- Auth middleware resolves token identity via userId instead of createdById
- New /api/user/api-tokens routes for self-service token management
- Admin /api/admin/api-tokens routes support userId and role overrides
- API Tokens section on profile page for all users
- Admin API tab shows all tokens with user/role selectors
Add persistent cross-ASIN "works" mapping and client-side deduplication to improve library matching. Introduces a Prisma migration and models (Work, WorkAsin) plus src/lib/services/works.service for persisting dedup groups, seeding ASINs at request time, and sibling lookup. Adds a deduplication utility (deduplicate-audiobooks) that normalizes titles/narrators, compares durations, and returns grouping metadata; API routes (search, author, series) now deduplicate results before enrichment and fire-and-forget persist groups. Adds sibling-ASIN expansion into audiobook matcher and expands getAvailableAsins accordingly. Extracts runtime parsing into a shared parse-runtime util and updates audible scrapers/services to use it. Includes unit tests for dedup logic and works service and updates test Prisma mocks.
Add support for hiding audiobooks that are already available by introducing a hideAvailable query flag and excluding matching ASINs at the DB level. Implemented getAvailableAsins() in audiobook-matcher to gather ASINs from the library and completed requests, and wired it into the popular and new-releases API routes to apply a notIn filter. Propagated the hideAvailable flag through useAudiobooks so client requests include the parameter, and adjusted the homepage to reset pagination when the flag changes. Replaced two StickyPagination instances with a new UnifiedPagination component (new file) that provides a single context-aware floating paginator which tracks the dominant section and allows switching between Popular and New Releases. Also removed client-side filtering in favor of server-side exclusion and made small imports/cleanup in page.tsx.
Add support for per-request custom search terms and an admin retry-download flow.
- DB/schema: add custom_search_terms column via Prisma migration and schema update.
- Admin UI: new AdjustSearchTermsModal component and UI badges to show custom search status; RequestActionsDropdown and RecentRequestsTable updated to surface adjust/retry actions.
- API: new PATCH /api/admin/requests/[id]/search-terms to set/clear custom terms (optionally trigger a new search) and new POST /api/admin/requests/[id]/retry-download to resume monitoring or re-add downloads using DownloadHistory metadata.
- Behavior: interactive search now prefers customSearchTerms when present; manual import exposes cleanupSource option to organize job; admin requests listing returns downloadAttempts and customSearchTerms.
- UX: add SectionToolbar, LoadMoreBar and HideAvailableToggle components and wire hide-available preference across home, search, author and series pages; authors/series endpoints/page handlers gain pagination metadata.
- Misc: add connection-errors util and update related processors/services and tests to cover the new flows.
These changes enable admins to override search terms per request, trigger searches from the admin UI, and retry failed downloads more robustly.
Introduce manual import workflow and download permission support. Adds a Prisma migration and schema field (users.download_access) to track per-user download access, and updates admin UI to toggle global and per-user download access. Implements new APIs: filesystem browse, manual-import endpoint, download-access settings, audiobook download-status, and on-demand download-token generation. Adds frontend components for manual import and related tests, plus documentation for the manual-import feature and the documentation-agent prompt. Key files: prisma/migrations/20260212000000_add_download_access_permission/migration.sql, prisma/schema.prisma, src/app/api/admin/filesystem/browse/route.ts, src/app/api/admin/manual-import/route.ts, src/app/api/admin/settings/download-access/route.ts, src/app/api/requests/[id]/download-token/route.ts, src/app/api/audiobooks/[asin]/download-status/route.ts, and updated admin users pages/components and permissions util.
- jwt.ts: Use JWT_DOWNLOAD_SECRET instead of JWT_SECRET for download tokens
- audio-formats.ts: Add EBOOK_EXTENSIONS export alongside AUDIO_EXTENSIONS
- request-statuses.ts: New shared COMPLETED_STATUSES constant used across requests API, download route, and RequestCard
- requests/route.ts: Import COMPLETED_STATUSES; strip filePath from audiobook in API response
- download/route.ts: Import format/status constants; add archiver dependency and replace adm-zip with streaming archiver for multi-file zips
- RequestCard.tsx: Use shared COMPLETED_STATUSES constant
Embeds a signed JWT download token (30-day expiry) in the requests API
response so users can download completed audiobook/ebook files directly
from the UI or by sharing the URL to apps like BookPlayer — no session
cookie required.
- jwt.ts: add generateDownloadToken / verifyDownloadToken helpers
- api/requests: append downloadUrl to completed requests with a filePath
- api/requests/[id]/download: new token-authenticated streaming endpoint;
serves single files directly or zips multi-file audiobooks with adm-zip
- RequestCard: add Download link in the actions area for completed requests
Always use qBittorrent's content_path as the canonical downloadPath and expose savePath on DownloadInfo instead of reconstructing paths from save_path + basename. Add path-waiting logic to the monitor: track consecutive pathWaitCount polls, re-queue the monitor with exponential-ish backoff while content_path remains outside save_path (to handle TempPathEnabled races), and give up after a configurable max attempts. Extend the MonitorDownload payload and JobQueue APIs to carry pathWaitCount. Organize-files processor now attempts to refresh the stored downloadPath from the download client and updates downloadHistory if the client reports a different path (applying path mapping). Update tests to reflect the new behavior and expectations.
When a torrent is finished (seeding/completed), build the download path from save_path combined with the basename of content_path instead of using torrent.name or the full content_path. This fixes a race with qBittorrent's TempPathEnabled (where content_path may still point to the temp dir) and addresses cases where the displayed torrent.name differs from the actual root folder/filename on disk. Added/updated tests to cover the TempPathEnabled race, name-mismatch scenarios, empty content_path fallback, and single-file torrents.
Add support for admin-driven interactive torrent selection and a file rename/template feature. Integrates an InteractiveTorrentSearchModal into the pending-approval admin UI, adds an admin approve flow that accepts an admin-selected torrent, and surfaces user/admin-selected torrent details in the UI. Introduces fileRenameEnabled and fileRenameTemplate settings (API + UI), persists them to configuration, and clears related caches. Pass renameConfig through the organize/organizeEbook flows and implement renaming in the FileOrganizer (single/multi-file handling). Enhance path-template utilities with conditional block resolution, filename-template validation, mock filename previews, and a buildRenamedFilename helper. Update tests to cover conditional templates and filename preview behavior.
Resolve downloadPath using save_path for finished torrents to avoid TempPathEnabled race conditions where content_path can point to a stale temp location. Compute status once, treat 'seeding'/'completed' as finished, and prefer path.join(save_path, name) for those states while still using content_path (or falling back to save_path) for active downloads. Added tests covering multiple qBittorrent states (seeding/stalledUP/pausedUP/stoppedUP/forcedUP/queuedUP/downloading and empty content_path) and imported path in tests.
Add DOWNLOAD_CLIENT_TIMEOUT (60000ms) in src/lib/constants/download-timeouts.ts and replace hardcoded 60000ms timeouts across Deluge, Prowlarr, qBittorrent and Transmission integrations. This centralizes the download/API timeout (gives headroom for indexers that enforce ~30s waits) and makes future adjustments easier without changing behavior.
Trim whitespace in the French stopWords array (add missing space after comma) to keep formatting consistent, and rename AUDIBLE_REGIONS.fr.name from "French" to "France" to better reflect the region label used for the Audible configuration.
kikootwo