Tankarr v0.10.0

Tankarr v0.10.0 consolidates the work since v0.9.0, including the reviewed community contributions and fixes for issues #24, #26, #41 and #52–60.

• Choose and remember a preferred download source for each series, with normal fallback when it cannot supply the requested content (#24).

• Optional OpenID Connect login using discovery, PKCE and explicit administrator allowlists; local login remains available (#26).

• English and Italian interface, manual catalogue entries, clearer chapter/volume coverage, improved large-library performance, API keys, notifications, system tasks/logs/metrics, and deployment examples for Helm and Unraid.

• Safer translation and language verification based on original-page evidence, with verified work preserved across interruptions. Strong release-level OCR evidence can resolve inconclusive artwork without overriding contradictory language evidence (#41).

• Pack Search panels close correctly and discard cancelled or stale requests (#52).

• A WEBTOON distribution link no longer overrides an explicitly printed series; Library, Wanted and acquisition agree on owned coverage (#53).

• Shutdown waits for import publication and owned direct downloads; manual torrent numbering survives a restart (#54, #55, #60).

• Downloads moved outside the managed client category are quarantined individually, while unrelated downloads continue (#56).

• Retrying failed SABnzbd history submits a new job and saves its new client ID (#57).

• Cleanup visits eligible older records beyond the first 200, and cancelled translation attempts cannot overwrite a newer retry (#58, #59).

• Provider search failures return generic messages without exposing exception details; healthy providers still return their results.

• Includes the community fixes for whole-volume and mixed packs, transient download retries, stable series units, Wanted recovery and credential redaction. CI explicitly installs Poppler for PDF integration tests (#61).

• Back up the configuration/database before upgrading. This version adds persisted source preferences, manual catalogue data, import assignments and verification metadata; restore the pre-upgrade backup when rolling back to an older version.

• Optional extras no longer contribute to automatic acquisition or main-story completion. Remove the retired TANKARR_SPECIAL_CHAPTERS_OUTSIDE_BOOKS setting if you previously configured it. Official season-opening episodes still count as main chapters.

• OIDC is disabled by default. Enabling it requires HTTPS issuer/callback configuration and an explicit subject or group allowlist. It grants administrator access, and logout ends the local Tankarr session. Signed mock-provider tests pass; live-provider verification is still outstanding.

• The managed Suwayomi pin remains v2.4.2366. Sources and extension repositories remain operator-configured.

• Upgrade instructions: Upgrading.

• Public release-PR CI: 2,613 backend tests passed; 167 frontend tests passed; 47 browser tests and 1 demo browser test passed.

• Ruff lint and format, production frontend build, i18n (1,647 strings), deployment files, dependency review, CodeQL, docs build and Docker build/smoke checks passed.

• Source-search error disclosure was corrected; CodeQL alert #43 is fixed, with regression coverage for sensitive details and preserved partial results.

• CI run covers the final release PR head.

• Backend lint, formatting and full pytest gate passed.

• Frontend tests, production build and translation-string checks passed.

• Regression tests cover the changed behavior.

• Public snapshot checked for installation-specific addresses, paths and local handoff files.

• Configuration and recovery documentation updated.

Docker image: ghcr.io/renzodef/tankarr:0.10.0 (linux/amd64, linux/arm64)

Upgrade notes: see the upgrade guide.

What's Changed

Other changes

• Bump the python group with 2 updates by @dependabot[bot] in #1
• Merge Dependabot updates automatically and declare Vite module types by @Renzodef in #7
• Bump the frontend group across 1 directory with 10 updates by @dependabot[bot] in #2
• Judge Dependabot groups per dependency by @Renzodef in #8
• Audit fixes: libarchive for RAR, ledger-based orphan sweep, deployment samples by @Renzodef in #9
• *arr parity and performance: API key, update check, URL base, faster renders, image attestations by @Renzodef in #10
• Import mixed volume and chapter packs with conservative numbering by @PACE-1 in #38
• Make two tests independent of the host that runs them by @PACE-1 in #42
• Mask encoded credentials in download failures and notifications by @PACE-1 in #48
• Suwayomi: store the unit of whole-book releases and read "Volume N VF" as a book by @PACE-1 in #44
• Retry a download whose source timed out while Suwayomi is healthy by @PACE-1 in #45
• Require an actual job before Wanted reports a source slot queued by @PACE-1 in #50
• Language audit: one page's script guess does not refuse a book by @PACE-1 in #43
• Download selector: keep the series unit it resolved from every release by @PACE-1 in #46
• Keep transient Suwayomi timeouts out of release blacklists by @PACE-1 in #49
• Install Poppler for backend PDF integration tests by @Renzodef in #61
• Release v0.10.0 by @Renzodef in #62

New Contributors

• @PACE-1 made their first contribution in #38

Full Changelog: v0.9.0...v0.10.0

Tankarr v0.9.0

The first public release of Tankarr: a self-hosted manga, manhwa and comics manager for the *arr stack, a "Sonarr for manga".

Docker image: ghcr.io/renzodef/tankarr:0.9.0 (also :latest), for linux/amd64 and linux/arm64.

• Follows the series you read with the familiar *arr profiles (all, future, existing, none), a Wanted list that says what was tried, a Calendar, an Activity queue and History.
• Identifies works through the MangaBaka catalogue, cross-checked with MangaUpdates, AniList, Kitsu and MyAnimeList, instead of trusting download sites' titles and numbering.
• Downloads from many channels: a managed Suwayomi engine with Mihon/Tachiyomi-compatible extensions from the repository you configure, Prowlarr indexers with qBittorrent and SABnzbd, the Internet Archive, and a local import folder.
• Keeps the language on each release and checks indexer downloads with OCR before they enter the library.
• Follows chapters or volumes, whichever a work actually offers, and can assemble chapters into books.
• Writes a clean, portable CBZ library with ComicInfo.xml, covers, atomic hash-checked imports, a recycle bin and automatic backups; reads it with its built-in reader or drives Komga and Stump.

Read the installation guide, then first steps. Tankarr never runs without a login: the first start creates one and prints it to the log.

latest and version tags follow stable releases. Changes under review can be tested with branch-<name> images from branches in this repository; main itself does not publish an image. See upgrading.

Docker image: ghcr.io/renzodef/tankarr:0.9.0 (linux/amd64, linux/arm64)

Upgrade notes: see the upgrade guide.

Full Changelog: https://github.com/Renzodef/tankarr/commits/v0.9.0