Media Tracker

A movie and TV tracker for TriliumNext. Your whole library lives in one JSON note under a library root you choose, so it is a single note you can back up, inspect, or hand-edit — and the library root itself becomes the tracker UI.

Replaces stremio-sync@beatlink, which this addon absorbs.

Setup

1. Install and enable the addon.
2. Open its Settings page and paste a TMDB API key (free, from themoviedb.org/settings/api). This powers search, posters, and episode lists.
3. Create a note to hold your library (anywhere in your tree), then pick it on the Library Root tab in Settings. This is required — every tracked title is created as a child of it.
4. Optionally set up Trakt and/or Stremio on the Import tab.

Picking a Library Root wires it up automatically: the note is converted to a render note pointing at the tracker widget and given a movie icon, so opening it in your tree shows the tracker itself. Choosing a different note reverts the old one back to a plain text note, and clearing the setting reverts it without selecting a replacement.

Using it

The widget has three tabs.

Library shows everything you track, with filters that compose:

• a search box that narrows by title as you type (case-insensitive substring)
• Type (All / Movies / TV), Status, Collection, and Genre dropdowns

Every option carries its own count, and each dropdown is scoped by the filters before it but never by itself — so selecting one option never zeroes out the others, and a count always tells you how many rows that choice would show. Status is colour-coded throughout: grey for planned, blue for watching, green for watched, red for dropped.

Each row has a status dropdown, a ★ rating box (0-10, blank for unrated), an Episodes toggle for shows, and a × to remove the title. A show's row shows its progress — episodes watched and seasons started — and expanding it shows episodes watched and seasons completed.

Refresh runs a housekeeping sweep over the whole library: it re-fetches metadata and posters from TMDB, backfills missing ids and episode counts, and recomputes every show's status from its episode progress. It never changes a rating and never un-watches an episode.

Details page

Clicking a title's name opens a full details page: large poster, overview, genres, runtime, status, rating, collections, and a scrollable cast list with photos and character names.

For shows, each season is a collapsible section listing every episode with its own summary, still image, air date, runtime, and TMDB rating — with a checkbox to mark it watched, so you can track progress while reading what each episode is. Season sections show their own watched count, and the header shows overall episode and season progress.

Bulk actions save clicking through a long show:

• Watch all episodes / Unwatch all on the show header, covering every aired episode
• Watch rest of season on each season, starting from the first unwatched episode so it fills forward from where you are (it reads Watch whole season when nothing is watched yet)
• Unwatch season to clear one season, and a "Season complete" note in place of the button when there is nothing left to watch

Each of these is a single request that applies the whole range in one write, so marking a 250-episode show watched is one note update rather than hundreds, and it cannot half-apply. Status updates automatically: all episodes watched becomes Watched, some becomes Watching, none becomes Planned.

Everything on this page is fetched from TMDB live rather than stored, so it needs a TMDB key and adds no weight to your database note.

Collections

Collections group titles into shared universes and work like tags — a title can belong to several at once. + Add to collection on any row opens a picker listing every collection you have with its current membership checked, so nothing has to be typed or remembered. When you have collection groups, the picker is split into a labelled section per group — Mood, Franchise, and so on — each edited separately. A field underneath creates a new collection; typing a name that already exists (in any casing) reuses it rather than making a near-duplicate, and a newly created one appears under Ungrouped until you file it into a group in Settings.

Each tick saves immediately — there is no separate save step to forget.

This is deliberately manual. TMDB cannot supply it: belongs_to_collection exists only on movies and covers narrow film series rather than universes (nothing there joins the MCU's films to its shows), and TV has no equivalent field at all. So Game of Thrones + House of the Dragon, or the MCU's films and series together, only group correctly when you say so.

Collection groups

Collections can be organised into groups — Mood, Franchise, Format, whatever you need — and each group gets its own titled dropdown on the Library tab. Selections across groups combine with AND, so picking Franchise: MCU and Mood: Comfort narrows to titles in both.

Define groups on the Collections tab in Settings. Each group gets its own block listing its collections, with a text box to add another directly into that group — so a collection is created where it belongs rather than being filed afterwards. A collection can still be moved between groups with its dropdown.

A collection created this way exists straight away, before any title uses it, and is marked unused until one does. Collections that belong to no group (created before groups existed, or left behind when a group was removed) appear in an Ungrouped block, which is omitted entirely when empty. Removing a group moves its collections there rather than deleting them.

Groups live in settings, not on the titles — a title still just carries collection names — so you can reorganise your groups freely without rewriting any title data.

Each group's dropdown lists its collections with counts, plus a None option showing titles that are in no collection of that group. This is per-group rather than global: Franchise: None finds titles you haven't assigned a franchise, even if they do have a Mood — which is what makes it useful for spotting gaps in one axis at a time.

Choosing one reveals ✎ (rename) and × (remove) beside it. Both sweep the whole library in one pass: rename updates every title carrying that collection, and remove clears the tag from all of them. Renaming onto a name that already exists merges the two rather than creating a near-duplicate.

Collections exist only as tags on titles — there is no separate list — so a collection disappears automatically once no title references it. Removing one never deletes any title; it only clears the tag.

A Group by collection toggle renders rows under collection headers instead. A title in several collections appears under each — groups overlap by design.

Genres

Genres are separate from collections and work differently: they come from TMDB automatically and are refreshed by Refresh, whereas collections are yours and never touched. They appear as their own dropdown with counts, scoped the same way — by type, search, and collection, but not by the genre filter itself.

Genres can be switched off entirely with Enable Genres on the Library settings tab. With it off, the Genre dropdown disappears, genres are hidden on the details page, and any leftover genre filter stops applying — useful if you would rather categorise everything with your own collections. Nothing is deleted, so turning it back on restores your genre setup as it was.

Because TMDB assigns a lot of genres, the Genres tab in Settings lists every genre in your library with a checkbox. Unticking one removes it from the filter row; it never changes a title's data. Show all / Hide all are there for a fast reset.

A title imported without TMDB metadata has no genres until you run Refresh. Note also that TMDB uses different vocabularies for film and TV — movies get "Science Fiction", shows get "Sci-Fi & Fantasy" — so both can appear as separate pills.

Sorting

Sort by A-Z, Recently watched, Release date, Rating, Recently added, or Progress, with an arrow button to flip direction. Titles missing the sort value (unrated, never watched, no release year) always sort last in both directions, so an empty field never leads the list.

Remembered view

Your status filter, type filter, collection filter, genre filter, sort field, sort direction, and grouping toggle are all saved, so the Library opens the way you left it. They live in the addon's settings note as hidden fields — persisted, but not shown on the Settings page, since the widget manages them.

The search box is deliberately not remembered: a text filter silently hiding most of your library on load reads as data loss rather than a convenience.

Add searches TMDB. The same All / Movies / TV chips scope the search: All uses TMDB's multi-search, while Movies and TV use the dedicated endpoints, which return a full page of one kind rather than a mixed page filtered down. Switching the chip re-runs the current search immediately.

Results you already track show a green ✓ Added marker instead of an Add button, so you can see at a glance what is new. It is matched on any shared id rather than only the TMDB id, so a title imported from Stremio (which supplies only IMDb ids) is still recognised.

You can also paste a link instead of searching. The button switches to Add and the title is added directly, no search step:

https://www.themoviedb.org/movie/693134
https://www.themoviedb.org/movie/693134-dune-part-two
https://www.themoviedb.org/tv/95396-severance/season/2
https://www.imdb.com/title/tt15239678/
tt15239678

A TMDB link names its own type in the path (/movie/ or /tv/), so it resolves in one call. The -slug suffix TMDB appends, trailing paths like /season/2, and query strings are all ignored — only the numeric id matters. An IMDb link or bare tt id is resolved through TMDB's /find endpoint.

Import is the one-way Trakt and Stremio import described below.

A Settings button at the end of the tab row opens the settings page, and a Back button there returns you to the tracker — to the library root you came from, or the addon's launcher note if no library root is set yet.

How titles are stored

Every tracked title lives in one JSON note titled Database, created automatically as a direct child of your Library Root. Keeping it under the root rather than inside the addon's own tree means the data travels with the library: move or export the root and your titles come along.

{
    "titles": {
        "tmdb:693134": {
            "tmdbId": "693134", "imdbId": "tt15239678",
            "mediaType": "movie", "title": "Dune: Part Two", "year": "2024",
            "status": "watched", "rating": 9, "lastWatched": "2026-07-20"
        },
        "tmdb:95396": {
            "tmdbId": "95396", "mediaType": "show", "title": "Severance",
            "status": "watching", "watchedEpisodes": "s01e01-e09,s02e01",
            "totalEpisodes": 19
        }
    }
}

Each entry is keyed by its strongest known id — TMDB, else IMDb, else Trakt. That key is what makes imports from different sources converge on one entry instead of duplicating: a title first imported from Stremio (IMDb id only) is still recognised when Trakt later supplies its TMDB id, because matching checks every shared id, not just the key.

The database note is a plain JSON code note, so you can open, inspect, back up, or hand-edit it. A blank, malformed, or partially-broken document is treated as empty rather than crashing the widget.

Since titles are JSON entries rather than notes, they are not individually linkable or cloneable, and Trilium collection views cannot browse them — the widget is the UI.

Episode tracking

Per-episode progress is a single compact string on the show's entry rather than a list of every episode:

"watchedEpisodes": "s01e01-e10,s02e01,s02e03-e05"

Consecutive episodes collapse into runs, so a fully-watched ten-season show stays a short string instead of a 250-element array.

The Episodes button on any show opens a season grid where you click individual episodes to toggle them. Watch status is recomputed automatically: no episodes is planned, some is watching, all aired episodes is watched.

Episode lists come from TMDB. Imported shows often arrive with only an IMDb id — Stremio supplies nothing else, and a Trakt entry has no TMDB id if metadata enrichment was off or its fetch failed — so the TMDB id is resolved from the IMDb id via TMDB's /find endpoint and written back to the entry, making the lookup a one-time cost rather than something that repeats on every open.

Import

Import is one-way and additive: external services are read, and your local edits are never pushed upstream. Imports only add and update — a title removed from Trakt or Stremio upstream is left untouched in Trilium, so an external change can never quietly delete your data.

The single exception is the deliberate per-entry Delete from Trakt action described under Comparing and clearing Trakt history, used when migrating off Trakt. Nothing is ever written to Stremio.

Trakt

Authorization uses Trakt's device flow: click Authorize, and the widget shows a code plus a URL. Enter the code there in your browser and the widget picks up the approval automatically. Tokens refresh automatically.

First create a Trakt API app at trakt.tv/oauth/applications, then copy two values from it into Settings:

• Client ID — the app's public identifier
• Client Secret — the separate secret value from the same app

If you get "Trakt does not recognise this Client ID", the usual causes are pasting the Client Secret into the Client ID field, or pasting only part of the value. Surrounding whitespace is trimmed automatically, so a stray newline from copying is not the problem.

A 403 Forbidden means something different: the request was blocked before reaching Trakt. Trakt requires a User-Agent header on every call and Cloudflare rejects requests without one, so this addon always sends one. If you still see a 403, check whether a proxy sits between your Trilium server and api.trakt.tv.

Trakt import brings in watched movies and full per-episode history for shows.

Stremio

Enter your Stremio email and password in Settings, then click Login on the Import tab. The password is discarded once an auth key is obtained.

Stremio only records your current position per show rather than a full history, so episodes up to that point are marked watched.

Migrating off Trakt

Archive everything on the Import tab is for leaving Trakt. Regular Import from Trakt only reads /sync/watched, which is aggregate state — it has no individual watch timestamps, no ratings, no watchlist, and no collection. Deleting your Trakt data after only that import loses all of it.

Archive fetches every sync endpoint, following pagination to the last page:

Endpoint	What it holds
/sync/watched/{movies,shows}	Aggregate watched state and play counts
/sync/history/{movies,episodes}	Every individual watch, with its watched_at timestamp
/sync/ratings/{movies,shows,seasons,episodes}	Your ratings
/sync/watchlist/{movies,shows}	What you planned to watch
/sync/collection/{movies,shows}	What you own

Trakt's raw, unmodified responses are saved to a Trakt Archive JSON note beside your library's Database note. That is the important part: even if the mapping into the library misses a field, the original data is still there to re-derive from later.

On top of that, watch data is imported into your library as usual, and Trakt ratings are applied to matching titles (existing ratings are kept unless you enable Let Imports Overwrite My Ratings).

Suggested order:

1. Click Archive everything and wait for it to finish.
2. Check the per-endpoint counts it reports against what Trakt shows you. If any endpoint says failed, fix that first — the report says explicitly not to delete yet.
3. Open the Trakt Archive note and confirm it looks complete.
4. Back up your Trilium data.
5. Only then remove your data on Trakt's own site.

Comparing and clearing Trakt history

Compare with Trakt fetches your full Trakt watch history and shows it against your Trilium library, newest first. Each row is marked:

• ✓ — that exact watch is recorded in Trilium (for an episode, that specific season/episode is marked watched; for a movie, the title is marked watched)
• ! — not in Trilium yet

Rows marked ! get an Import button that records just that one watch in Trilium — an episode is merged into its show's existing progress, never replacing it — after which the row flips to ✓ and can be deleted. This means you can work straight down the list without leaving the view.

Rows marked ✓ get a Delete from Trakt button that removes that single watch from Trakt. Rows marked ! cannot be deleted — the button is disabled and the backend refuses, so an unsaved watch can't be lost even by a crafted request.

Deletion targets Trakt's own history id, which is the correct way to remove one play: removing by title + timestamp is ambiguous, because Trakt does not guarantee that pair is unique, and could clear more plays than intended. There is no bulk delete, by design.

Trakt history removal is permanent — there is no undo on Trakt's side. Archive first, verify the counts, back up Trilium, then delete. The watch remains in Trilium either way.

This is the only place the addon writes to an external service; everything else is read-only, and Stremio is never written to at all.

Automatic imports

Both sources can import on a schedule. On the Import settings tab:

• Auto-Import From Stremio / Auto-Import From Trakt — enable per source
• Hours Between Auto-Imports — minimum gap between runs (default 6; use 1 for hourly, 24 for once a day)

A background script wakes hourly and imports only if that many hours have passed, so the interval is independent of Trilium's hourly-only scheduling. Each source needs to be logged in or authorized first; one failing doesn't stop the other, and a failure still records the timestamp so a broken source doesn't retry every hour forever.

The outcome of the last automatic run is shown at the top of the Import tab. Scheduled imports use exactly the same additive code path as the buttons — they only ever add and update.

Import safety

Repeated imports are safe and idempotent:

• Imports only ever add and update — they never delete. A title in your library that the source doesn't return is left completely untouched, so removing something from Trakt and re-importing does not remove it here, hand-added titles always survive, and an empty or failed response cannot wipe your library. The only thing that deletes a title is the × button.
• Titles are matched on any shared id (TMDB, IMDb, or Trakt), so two sources converge on one entry instead of creating duplicates.
• Episode progress is merged, never replaced — an episode you marked watched locally is never un-marked by an import.
• Your ratings and watch status are never overwritten unless you enable the matching Import setting.
• An import reads the database once, applies every change in memory, and writes once at the end, so a large import is a single note write and cannot half-apply if something fails partway.

Settings

Tab	Settings
Library	Library root note, status for new titles, poster size
TMDB	API key
Trakt	Client ID and secret; tokens are filled in automatically
Stremio	Email and password; auth key is filled in automatically
Import	Whether imports set watched status, fetch TMDB metadata, and may overwrite ratings

Migrating from stremio-sync@beatlink

The old addon wrote a flat HTML table into a note of your choosing; it kept no structured data, so there is nothing to migrate. Set up Stremio on the Import tab here and run an import to rebuild your library, then uninstall stremio-sync@beatlink. Its target note is left untouched — delete it yourself when you no longer want the old table.