v0.8.0

Films: the release where Transpondarr tracks both anime formats as themselves.
Add a film from AniList and it is searched on its title and release year,
grabbed, and filed into a Movies library of its own; episodes keep going where
they went before, and gain a flat filing option for the scanners that prefer it.
The REST resource series is also now titles, which API clients need to know
about.

Quality profiles are now chosen when you add a title rather than reassigned
afterwards, and the controls around them hold up — the chip survives a slow or
failed load, and names no longer collide only by case.

Added

• Anime films can be added, searched and grabbed. A film in AniList search
  or on the Discovery chart is no longer reserved and greyed out: adding one
  tracks it like any other title, and the same pipeline searches, grades, grabs
  and imports it. A film's releases are matched on its title and release year,
  which is what they are named for, so a manual search lists what is available
  and a monitored film joins the scheduled search queue and the recent-feed poll
  alongside your series, backing off the same way when nothing turns up. Format
  decides and episode count never does, so a one-episode OVA or special is a
  series and stays one. Two things to expect: automation waits for a release
  year rather than guessing at one — a film whose year AniList has not
  published yet is listed as ineligible until a metadata refresh fills it in,
  and a release naming a different year is listed with that mismatch as its
  reason — and a batch or season pack is held back for the same kind of reason,
  being as likely to be the parent series' pack as a multi-part release of the
  film. Searching and grabbing by hand work throughout, in both cases. Adds
  migration 00022.
• Films import into their own library directory. Plex and Jellyfin want a
  Movies library separate from Shows, so Settings → Library now takes a
  movies directory alongside the existing one, and a film places into it as
  Placeholder Film (2019)/Placeholder Film (2019).mkv, with the year left out
  of the name when the provider publishes none. Until that directory is set, a
  grabbed film waits in the Activity queue and says so rather than landing in
  the wrong library; setting one imports it on the next scan, and adding a film
  with no movies directory configured says so on the add form, so it is clear
  before the download finishes rather than after. A download that ships the
  feature alongside deleted scenes, an interview or a making-of imports the
  feature — the largest video in the payload. When two videos are the same size
  nothing is guessed: the import waits in the Activity queue for you to say
  which is the film.
• Films read as films across the web UI. A film's page carries a status card
  — wanted, downloading or in the library — with the release year in its header,
  and its add form asks only whether to monitor it. The library list reports what
  a film is doing: Downloading, Downloaded, not imported, Import
  blocked, In library, Wanted, or Not monitored when nothing is
  pursuing it. The calendar shows a film's premiere as a premiere, and places far
  more of them — the release date AniList publishes now fills in where a
  broadcast schedule never existed, so a dated film leaves the No schedule
  data footer and lands on its day, while one announced with only a year stays
  in the footer rather than being placed on a date nobody published. A film's
  Releases tab lists everything the search returned, and its Search button on
  the Wanted page links straight there. Wanted says Not released yet and
  states the date plainly; the Activity queue, the import notifications, the
  delete confirmation and the release-match reasons all name the film. Series
  screens keep their wording throughout, single-episode OVAs and specials
  included.
• Episodes can be filed without season folders. Settings → Library now
  takes a Series layout: keep the default season folders, or choose flat and
  have episodes land in Placeholder Saga/Placeholder Saga - S01E05.mkv for the
  scanner configurations that prefer it. Films are unaffected, and so are
  single-episode OVAs and specials, which stay series-shaped as before.

Changed

• A title's page has moved from #/series/:id to #/titles/:id, and the web
  UI calls it a title. The URL now matches what the rest of the product calls
  it. There is no redirect from the old address, so a saved link or a
  browser-history entry pointing at #/series/:id will not resolve — open the
  title from the Titles list instead. Every link inside the app already points at
  the new address. The sidebar and page title now read Titles to match, while
  genuinely episodic screens keep their wording.
• A title page's link parameters now do one job each. ?tab=releases opens
  the Releases tab and ?item=N filters it to one episode, where a single
  ?item=N used to do both. A saved link or bookmark carrying only ?item=N
  now opens the title on its usual landing tab rather than on Releases; a
  series' Search button on the Wanted page sends both, so it is unaffected.
• The REST resource is now titles, not series. An endpoint called
  series that returns films would be a wart the 1.0 stability promise freezes
  in place. Every /api/v1/series... route is now /api/v1/titles..., and the
  fields that name or reference one follow — in the webhook payload as well as
  the REST API, so the two agree. Discord and ntfy now announce Title added
  rather than "Series added". Nothing else changes — same behaviour, same data,
  same database, and your notification settings carry over untouched.
• Adding a title now opens a form for it, carrying both add-time
  decisions. The monitor choice was a single control above the whole result
  list, and the quality profile was not asked for at all — an added series took
  the default and had to be re-assigned from its own page afterwards. Clicking
  Add on a result now opens a form for that title with Monitor and
  Quality profile on it, defaulted for the common case, so a seasonal show is
  still one confirmation and a long-runner is narrowed before its first search.
  The global "Monitor on add" control and the Add · future only button
  annotation are gone, and Discovery opens the same form.

Fixed

• The quality-profile chip on a title's page no longer vanishes while it loads
  or when it fails. The picker rendered nothing until the profile list
  arrived, so the chips row shifted on every visit, and a failed fetch left no
  trace at all — just a missing control. It now holds its place while loading,
  shows a Profile unavailable chip you can click to try again, and links to
  Settings when there are no profiles to pick.
• Profile names no longer collide only by case. Saving a profile named
  Anime while anime exists is refused with the same conflict a duplicate name
  has always given, matching how group names within a profile are compared.

Internal

• Built with Go 1.26.6. The 1.26.5 standard library carries six
  vulnerabilities the release binary reaches — in net/http, crypto/tls,
  net/url, encoding/xml and encoding/asn1 — all fixed upstream.
• Duplicate-name detection reads SQLite's result code instead of matching on
  the error message, so a driver that rewords its errors can no longer turn a
  conflict into a 500.
• The profiles list reads groups and usage counts in bulk. GET /api/v1/profiles fetched both one profile at a time, so every profile cost a
  full scan of the series table. Groups and counts now come back in one query
  each, leaving the endpoint at three queries whatever the profile count.

Upgrade notes

• Set a movies directory before a film can import. Settings → Library
  takes one alongside the existing library directory (or
  TRANSPONDARR_LIBRARY_MOVIES_DIR); with it unset, films are still added,
  searched and grabbed, but a completed film waits in the Activity queue instead
  of being placed. Point it at a root your media server has as a Movies library,
  under the same mount your downloads use so the hardlink works. Nothing about
  episodes changes if you never set it.

• Saved links into a title's page need re-making. #/series/:id has no
  redirect to #/titles/:id, and a bookmark carrying only ?item=N now opens
  the title on its landing tab rather than on Releases. Links inside the app are
  already updated; only your own bookmarks and browser history are affected.

• The new series layout setting defaults to season folders, so nothing moves
  on upgrade. If you switch it later, it applies to future imports only:
  episodes already in your library stay where they are, and an upgrade of one of
  them lands in the new shape and leaves the old file behind, which the log
  warns about. API clients that write library settings must now send
  series_layout with the rest of the body: omitting it is refused rather than
  quietly resetting the layout to season folders. The web UI needs nothing.

• API clients using X-Api-Key need one search-and-replace; the web UI needs
  nothing. The UI ships inside the binary and is already updated, so only
  machine clients notice. Every route under /api/v1/series moved to
  /api/v1/titles — including {id}, {id}/search, {id}/grab,
  {id}/grabs, {id}/blocklist, {id}/blocklist/{entryId},
  {id}/pinned-group and {id}/profile — and these JSON fields were renamed:

  Before	After	Where
  series_id	title_id	activity queue and history, calendar, wanted groups, discovery entries, pinned-group and profile-assign responses
  series_title	title	activity queue and history, calendar, wanted groups
  series (list)	titles	GET /api/v1/titles
  series (name)	title	title search, grab history, title blocklist
  series (count)	titles	GET /api/v1/blocklist
  series_ids	title_ids	POST /api/v1/wanted/search
  series_queued	titles_queued	POST /api/v1/wanted/search, PATCH /api/v1/wanted/items
  series_count	title_count	GET /api/v1/profiles
  on_series_added	on_title_added	notification settings
  series_title	title	webhook payload
  event: "series_added"	event: "title_added"	webhook payload

  Generated clients also see renamed operation ids (list-series →
  list-titles, and so on) and schema names (SeriesDTO → TitleDTO,
  AddSeriesInputBody → AddTitleInputBody, and so on); regenerate against the
  new spec. Nothing is dropped alongside the renames — the series database
  table keeps its name and its columns, and saved notification toggles are read
  under their existing keys, so nothing needs reconfiguring.

  Films add fields rather than changing any: titles carry format and year,
  the calendar, wanted and activity rows carry format so a client can word a
  film as one, library settings take movies_dir, and adding a title accepts
  quality_profile_id. All are optional on the way in, so a client that ignores
  them behaves as it did.

  One rename fails quietly rather than loudly, so check it if you script your
  settings: a PUT /api/v1/settings/notifications that still sends
  on_series_added turns that notification off, because the toggles are
  plain booleans with no unset encoding and a save replaces all of them. Every
  other rename here reports itself — an old route 404s, and series_ids is a
  required field, so POST /api/v1/wanted/search returns 422 rather than
  resetting your whole library's search backoff.