Integrations¶
Tankarr runs without any other service: the built-in reader and the public
metadata catalogues work out of the box. Everything else on this page is
optional and is configured in Settings, one tab per area. Each panel has a
test button that checks the values in the form before you save them, and saved
values apply without a restart. Values saved in Settings are stored in
Tankarr's database and take precedence over the container environment.
Passwords and API keys are shown masked in Settings and are kept apart from the
other settings, in metadata.env in the data directory (/config in the
Docker image), readable only by its owner. Rarely needed fields appear after
turning on Show advanced.
Every setting also has a TANKARR_* environment variable. The
configuration reference lists them all with their defaults.
| Service | What Tankarr uses it for | Settings tab |
|---|---|---|
| Suwayomi | Chapters from manga sites, through Mihon/Tachiyomi-compatible extensions | Sources |
| Prowlarr | Searching your torrent and Usenet indexers | Indexers & torrents |
| qBittorrent | Downloading the torrent releases found through Prowlarr | Indexers & torrents |
| SABnzbd | Downloading the Usenet releases found through Prowlarr | Indexers & torrents |
| Internet Archive | Whole volumes downloaded directly from archive.org | Indexers & torrents |
| Readers | Opening the library in the built-in reader, Komga, Kavita, Stump or another reader | Reader |
| ntfy, webhook, Discord, Telegram, Apprise | Notifications | Notifications |
| Prometheus | Metrics scraping | API key |
| Metadata catalogues | Identifying works: titles, creators, covers, counts | Metadata |
| AI provider | Optional machine translation of missing books and chapters | Translation |
Suwayomi¶
Suwayomi-Server is an open-source server that runs Mihon/Tachiyomi-compatible source extensions. Tankarr uses it only as a source engine: through Suwayomi it searches the sources, reads their chapter lists and fetches the pages of a chapter. Monitoring, the download queue, CBZ packaging, the checks, naming and the import into the library stay with Tankarr. Tankarr does not use Suwayomi's own downloader or library updates; in managed mode they are switched off.
Configure it in Settings → Sources, panel Suwayomi. The variables are listed under Suwayomi in the configuration reference.
| Managed (default, recommended) | External | |
|---|---|---|
| Who runs Suwayomi | Tankarr, inside its own container, with the Java runtime included in the image | You, as a separate server |
| Setup | Install Suwayomi | Server URL, Username and Password |
| Extensions | Installed and updated by Tankarr, manageable from the panel | Whatever is installed on your server |
| Updates | Server and extensions checked once a day | Yours to manage |
Managed mode¶
Choose Managed by Tankarr (recommended) as the Mode and click
Install Suwayomi. Tankarr downloads the Suwayomi-Server release it was
tested with from GitHub, verifies the server JAR against both the SHA-256
digest pinned in Tankarr's own source and the checksum file published with the
release, installs it and starts it. Its data lives in the suwayomi folder of
the data directory. Tankarr supervises the process: it restarts it with a
back-off if it exits and stops it cleanly on shutdown. The panel shows its
status, the tested version, and offers Update / reinstall, Check
updates, Restart, Stop and the server Log.
Once a day Tankarr checks GitHub for a newer server release and reports it in
the panel and on the System page. It installs it on its own only when
Automatic server updates (TANKARR_SUWAYOMI_AUTO_UPDATE) is on and no
download is running; otherwise press Update to … when you decide. The
release pinned in Tankarr moves forward with Tankarr's own releases, so a
compromised upstream release cannot reach every installation within a day. The
same daily pass updates the installed extensions that come from the configured
repository and installs new extensions for your languages as described below;
Check updates does the same immediately. If an updated server does not
become ready, Tankarr restores the previous server and its data.
Java heap (MiB) (advanced; TANKARR_SUWAYOMI_MANAGED_HEAP_MB, default
256, from 128 to 1024) bounds the server's memory. With many extensions
installed, 512 is a safer value; raise it if the log shows
OutOfMemoryError. For the memory of the whole container, see
Requirements.
Network access and login¶
- With a Tankarr login, which is the normal case, the managed server listens
on port 4567 of the container's network interfaces, requires HTTP Basic
authentication with Tankarr's own user name and password, and serves its
bundled web interface. The credentials are written only to the server's
server.conf, readable by its owner only; they never appear on the Java command line, and the Java process does not inherit Tankarr'sTANKARR_*environment variables. - Without a Tankarr login (possible only with
TANKARR_AUTH_REQUIRED=false, on a development machine), the server listens only on the container's loopback interface, without authentication and without a web interface.
To open Suwayomi's own web interface, publish container port 4567 and use
Open Suwayomi in the panel. The button points to port 4568 on the host
name your browser uses for Tankarr, so either publish the port as 4568:4567
or set TANKARR_SUWAYOMI_PUBLIC_URL to the address you published. Other
containers on the same Docker network can reach port 4567 as well, and they
also need the Tankarr login.
Extensions¶
Tankarr ships no extension repository and recommends none. Enter the index
URL of a Mihon/Tachiyomi-compatible repository you trust in Extension
repository (TANKARR_SUWAYOMI_EXTENSION_STORE); the change is applied when
the managed server restarts, and until a repository is configured the panel
has nothing to offer and System shows an alert. The panel lists that
repository's extensions, plus any extension already installed. After
Install Suwayomi, and again during the daily maintenance, Tankarr installs
every extension of the repository for your Search languages (Settings →
General), except:
- extensions labelled NSFW;
- extensions the repository marks as obsolete;
- extensions for sources that answer only after a browser challenge (such as a Cloudflare challenge or a captcha) or that need an embedded browser.
When you have chosen your sources by hand (the Sources field is not empty), Tankarr leaves the installed extensions exactly as they are.
Tankarr never uninstalls an extension because it is slow, failing or obsolete. Reliability is handled by ranking, not by removal: see Source ranking and health.
The Extensions list in the panel shows the repository's extensions by language, with NSFW, Update available and Obsolete badges, and lets you install, update or remove any of them. An extension you install yourself is used whatever its content label. Test sources searches every enabled source for a few well-known titles and reports a verdict (good, slow when the average answer takes more than six seconds, or unreachable), the number of hits and the average answer time. The results also give new sources their first place in the ranking.
Auto-install official extensions (TANKARR_SUWAYOMI_AUTO_INSTALL_OFFICIAL,
on by default, managed mode only): when a work's catalogue record names a free
official platform, for example MANGA Plus or WEBTOON, Tankarr installs that
platform's extension and maps it as a source for the work.
Sources that are not supported¶
Sources that need a challenge solver or an embedded browser are not supported. The managed server runs without a challenge solver (its FlareSolverr support is switched off) and without Suwayomi's embedded browser, and Tankarr does not install extensions that depend on them. A source that answers with an anti-bot challenge is listed on System under Challenged sources and is ranked last.
Choosing sources¶
Click Test & load sources to list the installed sources for your search
languages, then choose in Sources which ones Tankarr may use. The automatic
selection (an empty field) uses every installed source in the language of the
series being searched, including sources installed later. A manual selection
stores Suwayomi's numeric source IDs (TANKARR_SUWAYOMI_SOURCE_IDS). A series
is only ever searched in the sources of its own language.
For a single series, open Download source on its series page, search the installed sources and choose Use this source on the matching title. Tankarr saves that source mapping and preference for this series, including after restart and metadata refresh. The chosen source takes priority when it offers a usable release; other eligible sources remain available when it cannot. A source in another language is refused, and normal numbering, language, ownership and import checks still apply. Existing library files are kept. Use automatic source selection removes the preference while retaining the verified mappings.
External mode¶
Choose External server as the Mode to use a Suwayomi server you already
run. Enter its Server URL, the service root without /api/graphql (for
example http://suwayomi:4567), and the Username and Password already
configured in Suwayomi, both or neither. Tankarr never creates or changes the
Suwayomi account. It uses whatever extensions are installed on that server;
install, update and extension management are available only in managed mode.
Prowlarr¶
Prowlarr manages your torrent and Usenet indexers. Through it Tankarr searches for whole volumes, packs and single chapters. It sends torrent results to qBittorrent and NZB results to SABnzbd, then imports the finished files itself.
Configure Prowlarr¶
In Settings → Indexers & torrents, panel Prowlarr indexers:
- Turn on Enabled and enter the Server URL: the address of Prowlarr as
Tankarr reaches it, without a path, for example
http://prowlarr:9696. - Paste the API key from Prowlarr → Settings → General → Security.
- Click Test & load indexers to load the indexers and their categories.
- Indexers: the automatic selection uses every indexer that is enabled in Prowlarr and advertises book, literature, manga or comic categories. You can pick indexers instead; indexers without such categories cannot be selected.
- Categories: only results in these categories are used, and at least one must stay selected. Select recommended picks the default: 7000 (Books), 7020 (EBooks) and 7030 (Comics).
The variables are listed under Prowlarr and Internet Archive.
How searches and grabs work¶
- Interactive Search on a series page, for one chapter, one book or the whole series, and the manual search on Wanted show indexer results next to the releases of your Suwayomi sources. Each result is labelled as a chapter match, a volume match or a series result and has a Grab button.
- The scheduled Wanted pass asks the indexers by itself for missing chapters, for the books that contain them, and for the missing volumes of series followed as books. It grabs a result on its own only when the match is unambiguous; anything else waits on System → To confirm. See Wanted recovery.
- The optional translation fallback also searches the indexers for editions in other languages, see Translation fallback.
For a grab, Tankarr resolves the exact result on the server and removes any API
key from its download link. For a torrent it checks that the magnet link or the
.torrent file carries the expected info-hash before handing it to qBittorrent.
The Prowlarr API key never reaches the browser.
Releases from indexers are imported only into series followed in English, after the OCR language check.
qBittorrent¶
qBittorrent downloads the torrent releases found through Prowlarr. Chapters from Suwayomi never go through it.
In Settings → Indexers & torrents, panel Torrent client (qBittorrent):
| Field | Variable | Default | Notes |
|---|---|---|---|
| qBittorrent URL | TANKARR_QBITTORRENT_URL |
unset | Web UI address as Tankarr reaches it, for example http://qbittorrent:8080 |
| Username, Password | TANKARR_QBITTORRENT_USERNAME, TANKARR_QBITTORRENT_PASSWORD |
unset | Web UI login; set both |
| Category (advanced) | TANKARR_QBITTORRENT_CATEGORY |
tankarr |
Tankarr treats every torrent in this category as its own |
| Automatic import | TANKARR_TORRENT_AUTO_IMPORT |
on | Import finished downloads once the numbering and language checks pass; when off, they wait in Activity for Import now |
| After import | TANKARR_TORRENT_COMPLETED_ACTION |
Keep seeding | See below |
| Orphan grace period (hours) (advanced) | TANKARR_TORRENT_ORPHAN_GRACE_HOURS |
24 | See below |
| qBittorrent link (browser) (advanced) | TANKARR_QBITTORRENT_PUBLIC_URL |
unset | Address a browser can open, for the shortcut on each torrent in Activity |
Test connection checks the address and the login.
Paths seen by the download client¶
qBittorrent usually runs in its own container, where the download folder has a different path than in Tankarr's container. Tankarr works with both, like the remote path mappings of Sonarr and Radarr:
TANKARR_QBITTORRENT_SAVE_PATH(default/data/downloads/tankarr) is the folder as qBittorrent sees it; Tankarr sends it with every torrent.TANKARR_TORRENT_DOWNLOAD_DIR(for example/downloads) is the same folder as Tankarr's container sees it. Without it, finished torrents cannot be imported.
Tankarr only reads these files, so a read-only mount is enough to import. With
a read-only mount, a download whose import is refused stays where it is, for
you to inspect. With a writable mount, Tankarr detaches a refused download from
qBittorrent without deleting it and moves its files into a
.tankarr-download-recycle folder on the download filesystem, where they are
kept for the recycle-bin retention period. See
Paths seen by the download clients
and Download clients for a complete
example.
Numbering in mixed packs¶
Tankarr preserves explicit volume and chapter markers, including (v02) and
[c001], when removing trailing release tags such as (2024) (Digital) [group].
For files with a bare number, a matching release title or containing folder such
as Series v01-17 + c148-224 (also Series v01-17 + 148-224) identifies the
volume and chapter ranges. Preview and import use the same rules.
A bare number beside explicitly numbered books is left unassigned when that
context is missing, conflicting or does not cover the number. Assign its volume
or chapter in the file picker. Packs containing only bare numbers retain their
volume numbering. An archive named Series 1 – 6 remains unassigned: split it
into individual books or deliberately assign it as an omnibus.
Library scans use the same folder hints and separate volumes and chapters into import groups for the same series. Unassigned files require a number before the import can start; their position in the list is not a book number.
After import¶
| After import | What happens to the torrent |
|---|---|
Keep seeding (seed, default) |
It stays in qBittorrent under qBittorrent's own ratio and time rules. |
Remove after import (remove_after_import) |
Tankarr deletes it and its files from qBittorrent as soon as the books are in the library. |
Remove when seeding is done (remove_when_seeded) |
Tankarr deletes it and its files once qBittorrent stops it at its seeding limits. |
About once an hour, Tankarr deletes, with their files, the torrents it added to its category whose download was deleted afterwards (a series removed, a chapter unmonitored) and that are older than the orphan grace period. Tankarr keeps a record of every torrent it hands to qBittorrent and only ever removes those: a torrent another application or you put in the same category is left alone and listed on the System page. Torrents outside the category are never touched.
Moving a managed download to another client category marks that download as failed without interrupting other downloads. Restore Tankarr's category before using Retry. Retention cleanup checks all retained imports in pages, so a long history of completed downloads does not hide older seeding jobs.
A magnet link whose metadata no peer delivers within 30 minutes
(TANKARR_TORRENT_METADATA_TIMEOUT_MINUTES) is removed from qBittorrent and its
download fails, so the next candidate release gets its turn.
In Activity, Remove release and staging files removes a release and its files from the download client; books already imported stay in the library.
SABnzbd¶
SABnzbd downloads the Usenet (NZB) releases found through Prowlarr. In Settings → Indexers & torrents, panel Usenet client (SABnzbd):
| Field | Variable | Default | Notes |
|---|---|---|---|
| SABnzbd URL | TANKARR_SABNZBD_URL |
unset | Address as Tankarr reaches it, for example http://sabnzbd:8080 |
| API key | TANKARR_SABNZBD_API_KEY |
unset | SABnzbd → Config → General → API Key |
| Category (advanced) | TANKARR_SABNZBD_CATEGORY |
tankarr |
Tankarr treats every download in this category as its own |
| Completed folder (as SABnzbd sees it) (advanced) | TANKARR_SABNZBD_COMPLETE_PATH |
/data/downloads/usenet |
SABnzbd's completed-downloads folder, as SABnzbd sees it |
| SABnzbd link (browser) (advanced) | TANKARR_SABNZBD_PUBLIC_URL |
unset | Address a browser can open, for the shortcut in Activity |
TANKARR_USENET_DOWNLOAD_DIR (for example /usenet) is the completed folder as
Tankarr's container sees it; without it, finished Usenet downloads cannot be
imported. A read-only mount is enough. Test connection checks the address
and the key, and creates the category in SABnzbd when it is missing.
Automatic import and After import in the qBittorrent panel apply to Usenet downloads too. With Keep seeding, a finished download stays in SABnzbd's history; with either removal option, Tankarr deletes it and its files from SABnzbd right after the import.
Retry resubmits a failed NZB even while its failed entry remains in SABnzbd's history. A download still queued or successfully completed is reused instead.
Internet Archive¶
archive.org holds scanned and digital comics as whole
volumes in CBZ, CBR and PDF. Tankarr uses it as a direct-download book source,
with no download client: it searches archive.org by the work's title (as an
exact phrase, among text items), reads each item's file list and downloads the
book file itself into the direct-downloads folder of the data directory. The
file then goes through the same import path as a finished torrent: identity
review, OCR language check, numbering and page quality. The downloaded copy is
removed once the book is in the library.
- Only the files the uploader provided are considered, not archive.org's own converted copies, and only files between 1 MB and 2 GiB.
- Tankarr asks archive.org at most once per second, once per Wanted pass for a series, and only for whole books, never chapter by chapter.
- Results rank below every indexer result, so archive.org is the fallback for books no indexer carries, typically out-of-print works.
- Like indexer releases, archive.org books are imported only into series followed in English.
It is on by default. In Settings → Indexers & torrents, panel
Internet Archive (direct download), Enabled
(TANKARR_INTERNET_ARCHIVE_ENABLED) includes archive.org in the series release
search and the Wanted book search, and Test archive.org runs a sample search.
Readers¶
Tankarr writes a plain folder of CBZ files, so any reader that opens such a folder works with it. Settings → Reader decides which reader the Open and Read links in Tankarr point to, and whether Tankarr keeps that reader's library in step.
| Reader | What Tankarr does |
|---|---|
| Tankarr (built-in) | Opens the CBZ files in its own web reader. Nothing to configure. |
| Automatic (default) | Komga when the Komga link is enabled, the URL template when one is set, otherwise the built-in reader. |
| Komga | Links to Komga and, with an API key, manages its library: scans, titles and covers. |
| Kavita | Links to the matching Kavita series. Kavita scans the folder on its own schedule. |
| Stump | Links to Stump and manages its library: scans and covers. |
| URL template | Builds a series link from a template. |
| No shortcut | No reader links. |
Default reading mode (TANKARR_READER_DISPLAY_MODE) sets how the built-in
reader shows pages: Automatic, Manga (right to left) or
Webtoon (vertical). See
Built-in reader and bookmarks.
Discover reader tries the reader addresses already configured and the usual
Docker service names (komga:25600, kavita:5000, stump:10801) and fills in
the fields for the reader it finds; it never scans your network, and nothing is
saved until you save. Test reader connection checks the current values.
When Tankarr cannot reach the browser address from its container, set
Reader URL from inside Tankarr (advanced) to the service name, for example
http://komga:25600.
Komga¶
Enter the Komga URL (the address your browser uses) and a Komga API key from Komga → Account → API keys. With both, Tankarr manages the Komga library automatically:
- it asks Komga to scan after the download queue drains, and runs a full
safety-net scan every 360 minutes (
TANKARR_KOMGA_REFRESH_INTERVAL_MINUTES, from 15 to 10080); concurrent requests are merged into one scan; - it publishes canonical titles, metadata and the selected covers to Komga;
- it empties Komga's trash of the records its own renames and deletions left behind, only once the library is proven mounted and complete, and it leaves the trash untouched if a single record in it cannot be explained;
- it never changes reading progress.
The API key is sent in the X-API-Key header. Komga can also be managed with
an administrator's user name and password over HTTP Basic, configured through
the environment: TANKARR_KOMGA_LINK_ENABLED=true, TANKARR_KOMGA_URL,
TANKARR_KOMGA_AUTH_METHOD=basic, TANKARR_KOMGA_USERNAME and
TANKARR_KOMGA_PASSWORD. Changing a setting in the Reader panel afterwards
replaces that setup with the panel's values. See
Reader, Komga and Stump.
Stump¶
Enter the Stump URL, a Stump username and Stump password, and in
Library path inside Stump (advanced, default /data/comics) the path at
which Stump's container sees the library folder. Tankarr finds series and books
by their exact folder and file paths, never by title. It asks Stump to scan
after the download queue drains and on the same safety-net interval as Komga,
and uploads the selected covers.
Kavita¶
Enter the Reader URL and a Kavita API key from Kavita → Settings → Account → API key. Tankarr uses the key to find the Kavita series that holds a Tankarr series' files, matched by their paths in the library, and links to it. It does not start Kavita scans.
URL template¶
Series URL template builds a series link from a template, for example
https://reader.example.com/search?q={title}. The placeholders are {title},
{title_raw}, {folder} and {id}.
Notifications¶
Tankarr sends the same three events to every channel you configure under Settings → Notifications, each panel with its own Send test notification button:
| Panel | Fields | Variables |
|---|---|---|
| Events | On chapter imported, On download failed, On decision needed | TANKARR_NTFY_ON_CHAPTER_IMPORTED, TANKARR_NTFY_ON_DOWNLOAD_FAILED, TANKARR_NTFY_ON_DECISION_NEEDED |
| ntfy | ntfy URL, Topic | TANKARR_NTFY_URL, TANKARR_NTFY_TOPIC |
| Webhook | URL, Bearer token | TANKARR_WEBHOOK_URL, TANKARR_WEBHOOK_TOKEN |
| Discord | Webhook URL | TANKARR_DISCORD_WEBHOOK_URL |
| Telegram | Bot token, Chat ID | TANKARR_TELEGRAM_BOT_TOKEN, TANKARR_TELEGRAM_CHAT_ID |
| Apprise | Server URL, Configuration key, Notification URLs | TANKARR_APPRISE_URL, TANKARR_APPRISE_KEY, TANKARR_APPRISE_URLS |
ntfy is active when both the server root and the topic are
set; Tankarr publishes without credentials, so use a topic that accepts
anonymous publishing. The webhook receives one JSON document per event, with
event, title, message, priority, tags, at and event data such as
the series and chapter, for Home Assistant, n8n or a script of your own.
Discord gets an embed, red for failures. Telegram needs a bot token from
@BotFather and the chat ID the bot posts to. An
Apprise API server forwards to most
other services, through a stored configuration key or the notification URLs
you give Tankarr. The three events are:
- Chapter imported: a verified file was written to the library (normal priority).
- Download failed: a download job reached Failed (high priority).
- Decision needed: new matches waiting on System → To confirm, and Wanted items that every channel has given up on ("Not obtainable"). Each item is announced once (normal priority).
Send test notification publishes a real test message with the values currently in the form, even before you save them. A failed notification never fails the import or the job that triggered it.
Monitoring (Prometheus)¶
GET /metrics exposes the library and queue counts, the free space of the
data and library volumes, the last run and state of every scheduled task and
whether an update is available, in the Prometheus text format. It answers
behind the normal login: give Prometheus the API key from Settings →
Security as a bearer token, or the login as basic auth.
scrape_configs:
- job_name: tankarr
scrape_interval: 60s
static_configs:
- targets: ["tankarr:8787"]
authorization:
credentials: <the API key>
Behind a URL base the path is /<base>/metrics. A scrape reads a few SQL
aggregates and the state the background loops already keep; it never renders
a page, so it costs milliseconds.
Metadata catalogues¶
Tankarr identifies works through public catalogues that need no account and no
API key. Settings → Metadata, panel Metadata catalogues, shows
MangaBaka · Always on with a Test connection button, and the
Refresh interval (hours) (advanced; TANKARR_METADATA_REFRESH_INTERVAL_HOURS,
default 168, one week).
- MangaBaka is the catalogue spine. Add New
searches it, and each series' record gives titles, creators, description,
cover, type, status, counts, official platform links and the identifiers of
the same work on the other catalogues. Author work lists are refreshed once a
day (
TANKARR_AUTHOR_REFRESH_INTERVAL_HOURS, default 24). - MangaUpdates, AniList, Kitsu and MyAnimeList, the latter through the public Jikan API, are read only by the identifiers MangaBaka provides, never searched by title. Their counts and statuses corroborate MangaBaka's, see Catalogue consensus. A catalogue that does not answer only removes one opinion.
System → Enrich Metadata refreshes every series immediately. See How Tankarr works for how the catalogue decides identity, titles and covers.
Translation fallback¶
When no release exists in a series' language, Tankarr can machine-translate a book or chapter from another language. It is off by default, both globally and for each series. Configure Settings → Translation (an OpenAI-compatible AI provider API base URL, AI model and AI provider API key, and optionally an external processor under advanced settings), then enable the fallback in each series' Edit dialog. OCR and lettering run on the Tankarr host by default; the AI provider receives the recognized dialogue text and may charge for it. The variables are listed under Translation, and Translation fallback explains the whole process.
External services Tankarr contacts¶
Besides the services you configure yourself (Prowlarr and, through it, your
indexers; qBittorrent; SABnzbd; your readers; your ntfy server; your AI
provider or translation processor), Tankarr contacts only these public
services. Its requests identify themselves with the user agent
Tankarr/<version> (+https://github.com/Renzodef/tankarr).
| Service | Address | When |
|---|---|---|
| MangaBaka | api.mangabaka.org |
Searches in Add New, series and author records, metadata refreshes |
| MangaUpdates | api.mangaupdates.com |
Metadata refreshes, by catalogue ID |
| AniList | graphql.anilist.co |
Metadata refreshes, by catalogue ID |
| Kitsu | kitsu.io |
Metadata refreshes, by catalogue ID |
| Jikan (MyAnimeList data) | api.jikan.moe |
Metadata refreshes, by catalogue ID |
| Catalogue image hosts | The image servers of MangaBaka and of the catalogues its records link to (AniList, MangaUpdates, Kitsu, MyAnimeList, Anime-Planet, Anime News Network) | Downloading the covers named in catalogue records; other hosts are refused |
| NAVER Webtoon | comic.naver.com |
Reading the official episode list of a work whose catalogue record links there, as numbering evidence only; nothing is downloaded from it |
| Internet Archive | archive.org |
When enabled: searches, item file lists and book downloads |
| GitHub | api.github.com, github.com |
Managed Suwayomi only: checking for and downloading Suwayomi-Server releases (at install and once a day), and the extension repository's index and packages |
The websites behind your Suwayomi extensions are contacted by the Suwayomi server, managed or external, when it searches and downloads for Tankarr. Discover reader contacts only the reader addresses you configured and the Docker service names listed under Readers.