Your library
Libraries
The collection can be split into named libraries — say Comics and Manga — each with its own entry in the sidebar. Create them in Settings → Library → Libraries: a library has a name, a type that sets how its series behave (manga = chapter-style search and right-to-left reading defaults), and an optional folder — new downloads for that library file there instead of the default root.
- Move a series between libraries from its ⋯ menu; it takes on the library's type.
- Import scans each library's folder too, and anything found under one joins that library automatically, typed correctly.
- A library can have its own folder pattern (e.g. just
{series}for a manga tree without publisher folders) — blank uses the global pattern. - Mark a library Mature to hide it — name, entry, and every series in it — from roles without the "View mature content" permission (e.g. a kids' account). Series moved in inherit the flag; moved out, they shed it.
- The arrows on each library set the order they appear in the sidebar.
- One comic or manga library is the default: series you add without choosing a library, and series matched from an import outside every library folder, land there. Pick it with Make default; until you do, it's the first comic library in the list.
- Deleting a library keeps all its series (they return to the default library) — nothing is removed from disk.
- With no libraries defined, the sidebar shows per-type entries automatically once a second type (e.g. manga) appears in the collection — explicit libraries simply take over when you create them.
Libraries
Split the collection into named libraries — each shows as its own entry in the sidebar. A library's type sets how its series behave (manga = chapter-style search, right-to-left reading); its folders are where its comics are filed and scanned. Move series from a volume's ⋯ menu.
{series} folder pattern; Comics leaves it blank and follows the global one. A library's type decides how its contents behave. Comics and manga follow the ComicVine flow described here; Books and Audiobooks are self-described libraries with their own scanning, metadata and reading/listening — see Books and Audiobooks.
Manga metadata and covers come from MangaDex, enriched by AniList and MangaUpdates. MangaDex identifies the series and provides the cover, alternative titles, status, genres and the kind of book (manga, manhwa, manhua, webtoon); the series' own AniList and MangaUpdates ids are then followed for a curated summary, staff, end year and the publisher. Publication status and genres show on the series page and drive the Ongoing and Ended filters, as they do for comics. The Add dialog always has Comics and Manga tabs; if you have no Manga library yet, one is created on the first manga add. Imports into manga folders match against the manga catalog automatically.
Storage locations
Storage locations live on your libraries (above): each library's folder is where its comics are filed — local paths or network shares (\\NAS\comics, /mnt/comics) — and where scans look for what you own. Upgrading from an older version migrates automatically: your former default scan folder becomes a Comics library, and any extra scan folders each become a library of their own.
Importing an existing collection
Coming from another collection manager, a hand-organized folder tree, or a pile of loose files? Sidebar → Import:
Coming from Mylar3 or Kapowarr?
The Migration Assistant reads their database directly and matches every series by its ComicVine volume id, which is exact where a folder-name scan can only guess. Do that first, then use Import for anything it leaves.
- Press Scan for new to walk your library folders for series not yet in the collection (it keeps your review work), or Full rescan to clear the list and start over.
- BackIssue walks it and proposes a match for each series folder against ComicVine. Tagged libraries match best: when the embedded
ComicInfo.xml(Mylar, ComicTagger, Kapowarr) carries a ComicVine id, the volume is matched exactly — otherwise the tagged series name, start year, and publisher drive the search; untagged files are matched from their folder names. - Strong matches are marked Ready; ambiguous ones become candidates under Needs review that you confirm or re-pick with a couple of clicks; anything unrecognizable is listed for manual handling. Nothing is imported until you press Import.
- Imported files are indexed as owned — the series' missing counts update immediately.
Scan your root folders for series not yet in the collection. Each is matched against the metadata source — confirm or fix the matches, then import. Files stay where they are.
Import never moves or renames your files unless you later run the rename tool.
Supported layouts: a flat Series/ tree, the default Publisher/Series (Year)/, and Mylar-style Publisher/Series/Volume/ trees (e.g. Marvel/X-Men/v2004) — when a folder is only a volume marker (v2004, Vol. 3 (1999)), the series name is read from the folder above it, and each volume folder is matched to its own ComicVine volume.
How files are recognized
The scanner reads real file contents, not just names:
- Archive type is sniffed from magic bytes — a RAR misnamed
.cbz(or vice versa) is still read correctly. - Embedded
ComicInfo.xmlmetadata is used when present and sane. - Filenames are parsed for series / issue number / year as a fallback, with the same strict matcher used for downloads.
- Files linked to a ComicVine issue update the owned/missing math; unlinkable files are flagged rather than guessed at.
Tagging and naming
- With Tag on download on (Settings → Metadata; off by default), every downloaded issue gets ComicVine metadata as
ComicInfo.xml(series, number, title, date, summary, creators, characters, teams, locations, story arc and page count) — the standard read by comic readers and library servers. Characters and arcs come from ComicVine where it has them and from Metron otherwise, which for anything published before about 1980 is nearly always. - Tag placement decides where that XML lives. Embedded (the default) writes it into the archive. Sidecar writes it to a
.xmlfile next to the archive — the archive itself is never modified, so its bytes stay identical for torrent seeding and file-share hashing, and.cbrfiles stay.cbr. Set it globally in Settings → Metadata, or per library (a seeding library can use sidecars while the rest embed). When a file has both, the sidecar wins. - Folder layout and filenames follow your naming patterns (below). The defaults —
Publisher/Series (Year)folders andSeries VYYYY #NNNfilenames — are unambiguous, sortable, and parseable by other library managers. Sidecar files are renamed, moved, and deleted together with their archive. - A series in a manga library is tagged as a volume:
Volumecarries the volume number,Numberis left out, andMangais set toYesAndRightToLeft, which is what Kavita and Komga expect. Comics keepNumberandVolume= the series' start year. - With embedded placement, CBRs are converted to CBZ on the way in (solid-archive-safe, pages stored without recompression), because CBZ is what embedding and readers handle best. Sidecar placement skips the conversion — originals are left untouched.
Existing files you imported keep their names until you opt into renaming (below).
Naming patterns
Settings → Library → File organization — you decide how comics are organized on disk, with two token patterns and a live example that updates as you type:
- Folder pattern — the series folder under a root. Default:
{publisher}/{series} ({year}). Use/for sub-folders. - File pattern — the issue filename. Default:
{series} V{year} {edition} #{issue}—{edition}drops out when the issue has no edition. A series in a manga library defaults to{series} Vol. {issue:2}instead ([Oshi No Ko] Vol. 01.cbz), because manga readers readV2023as volume 2023; a pattern you type applies to every library.
File organization
How downloaded comics are named and filed.
DC Comics/Batman (2011)/Batman V2011 #001 (11-2011).cbz 2
Changing these affects new downloads — for existing files use Reorganize library on the Tools page.
| Token | Fills with |
|---|---|
{publisher} | Publisher name |
{series} | Series title (any trailing year marker removed) |
{year} | The volume's start year |
{issue} | Issue number, zero-padded to 3 — {issue:2} sets the width |
{issueTitle} | The issue's title |
{date} | Cover date as "November 2011" — see the formats below |
{edition} | Detected special editions (Annual, TPB, …) |
Date formats
{date} writes the cover date as "November 2011". A modifier gives you the parts on their own, for libraries that file by number:
| Token | Renders |
|---|---|
{date} | November 2011 |
{date:m} | 11 — month number, always two digits |
{date:y} | 2011 |
{date:mon} | Nov |
Combine them with whatever separator you use. {series} V{year} #{issue} ({date:m}-{date:y}) gives:
Batman V2011 #001 (11-2011).cbzand ({date:y}-{date:m}) gives (2011-11), ({date:mon} {date:y}) gives (Nov 2011). The month keeps its leading zero, so March files as 03, not 3, and sorts correctly beside the rest of the year.
A modifier the app doesn't recognise falls back to the full "November 2011" form rather than rendering nothing, so a typo cannot quietly strip the date out of every filename.
When a token is empty
A token with no value simply drops out and the spacing tidies itself — {series} V{year} with no known year renders as just the series. Blank patterns use the defaults.
This extends to the punctuation around it. An issue with no cover date renders ({date:m}-{date:y}) as nothing at all, rather than leaving an orphaned (-) in the filename. Downloads often have no cover date at the moment they are filed, which is why the default file pattern carries no date token at all — a freshly downloaded issue and the same issue after a re-file would otherwise disagree, and the reorganizer would churn.
Two things to know:
- Changing patterns affects new downloads; existing files stay put until you explicitly reorganize (below).
- Rename downloaded files (a toggle in the same section, on by default): turn it off and completed downloads keep the source's original release filename, still filed into the comic's folder.
Applying patterns to existing files
Always explicit, never automatic:
- One series — the volume page's ⋯ → Rename files moves/renames that series' files to match your patterns (a confirm shows the count first).
- Whole library — System → Tools → Reorganize library. Preview first: a dry run shows what would move, grouped by destination folder, with counts for already-matching files and name collisions. Reorganize then runs as a background job with live progress (also visible under System → Jobs) — a big library on a NAS takes a while, and the app stays fully usable during it.
Safety rules for both: only ComicVine-matched series are touched, name collisions are skipped (never overwritten), files are filed under the series' own library folder (so a series you moved to another library has its files moved on the next rename, which the preview shows first), and emptied folders are cleaned up.
Library tools
System → Tools — library-wide maintenance, each with live progress:
| Tool | What it does |
|---|---|
| Scan entire library | Re-walk every library folder: index new files, drop records of deleted ones |
| Re-index series folders | For every matched series, re-index its own folder and attribute the files there — without fuzzy matching. Fixes files attached to the wrong same-named series (finding brand-new comics is still Scan entire library) |
| Tag all untagged files | Embed ComicVine metadata into every owned file that lacks it (converts CBR as needed) |
| Convert all CBR → CBZ | Repack every .cbr so the whole library is consistently taggable |
| Re-download corrupt files | Runs Unwrap nested archives first, then queues a fresh download of every other corrupt comic file. Each corrupt file stays until a good copy replaces it; files not matched to an issue can't be fetched and are only counted |
| Unwrap nested archives | Fix comics packaged as a .cbz that holds a .cbr instead of pages: lifts the inner archive's pages to the top level, in place, keeping the ComicInfo. Anything it cannot prove is left untouched |
| Remove duplicate files | Delete corrupt copies that a good copy of the same issue has replaced. Where two good copies exist, the best is kept — tagged first, then the most pages, then the largest — and the other removed; that part only previews (see Logs → tools for the list) unless you tick the box |
| Verify archives | Deep-check every file for corruption; prune records for files gone from disk |
| Refresh series metadata | Re-pull every matched series' volume details and issue list — picks up publication status (Ongoing / Ended), enrichment for series cached before it was on, and issues published since. One request a second; stops cleanly if the service rate-limits (run again to finish) |
| Re-link files to metadata | Re-map owned files to the metadata source's issues across the library (fixes owned/missing counts after big changes) |
| Merge duplicate series | Fold together series rows that point at the same volume. The folder-backed row is kept; the other's files, wanted issues and follows move over, and any duplicate copies that surface are cleared with Remove duplicate files |
| Re-check mismatched volumes | Find series whose files go beyond their matched volume's last issue — the sign a same-name mini or one-shot was matched instead of the real run — and re-match them. A clear winner is applied and its files re-linked; anything unclear is left for Fix match on the series page |
| Download issue metadata | Fetch ComicVine detail (descriptions, credits, dates, covers) for every issue in your collection that's missing it — already-cached issues are skipped, and it stops cleanly if ComicVine rate-limits (re-run to finish) |
| Rename files to pattern | Rename every CV-linked file to your file pattern, in place (same folder) — collisions are skipped, nothing is overwritten |
| Reorganize library | Move and rename every matched series' files to your folder + file patterns — dry-run preview first, then a background job with progress. See Applying patterns to existing files |
| Back up database | Snapshot catalog.db into backups/ (keeps the newest 5); safe while the app runs |
| Remove ghost series | Find leftover series with no match and no files on disk, whose wanted issues clutter the Wanted page but can never download. Only previews (see Logs → tools for the list) unless you tick the box |
Restoring a backup: stop the app, copy the snapshot over catalog.db, start again.
Stats
Sidebar → Stats is the read-only overview of what you actually have. Anyone who can browse the library can open it, and it answers in one page the questions that otherwise need a lot of clicking:
- What it weighs. Total size on disk, how many files are indexed, how many are valid versus corrupt, and how many carry embedded tags.
- Format mix. The split between CBZ, CBR, PDF and everything else — the quickest way to see whether a conversion pass is worth running.
- By publisher. Series, issues, files and size per publisher, largest first.
- Completion. How many series are complete, how many have holes, and how many issues are missing overall. Biggest gaps names the dozen series missing the most, which is usually where a backfill should start.
- Metadata health. How many series are matched, how many files are linked to an issue, and how deep the metadata cache runs. Unmatched or unlinked counts climbing is the early warning that a scan or re-link is due.
- Downloads. What is in progress, imported and failed, activity over the last fortnight, and the most recent imports.
Figures are cached for a minute, so the page is cheap to leave open.
One thing to know before you reconcile numbers: for a role that cannot see mature series, those series are left out of Biggest gaps and the recent imports, but the headline totals are not filtered. The totals can therefore exceed what that account is able to browse.
Corrupt files
Verification flags unreadable archives as corrupt (visible as a series badge and under the Problems filter). Redownloading an issue replaces the bad file; the duplicate-removal tool cleans up superseded bad copies afterwards.