Xtream Tuner

An Emby Server plugin that turns any Xtream-compatible IPTV service into a full Live TV, Movies, and Series library — with EPG, metadata matching, and a built-in dashboard.

---

This is a fork of firestaerter3/emby-xtream. It tracks upstream, adding a companion interface that de-duplicates titles across categories and tracks reviewed/unreviewed items to make processing large VOD libraries more efficient. It's currently most useful if you use Dispatcharr as your XC provider: Dispatcharr normalizes titles across providers before they reach Emby, giving movies a single stream ID and series an identical name across providers. Those are the two factors that drive this fork's de-duplication. While this approach is unlikely to help de-duplicate libraries imported directly from IPTV providers, the reviewed/unreviewed aspect of the companion interface will function regardless. Extending de-duplication to raw provider libraries is possible future work. Everything in the upstream README below still applies.

What this fork adds

When a provider lists the same title under many categories, a simple per-category view means reviewing the same movie or show multiple times. This fork adds a title-level review workflow on top of the existing per-title exclusion:

• De-duplicated review view — one row per unique title across your selected categories (movies collapse by stream ID, series by name), so a title listed in five categories appears once. Search, filter by category, and exclude at the title level.
• Reviewed checkpoint — mark titles "reviewed" (a bookmark, separate from excluding), with segmented Show (All / Included / Excluded) and Reviewed (All / Reviewed / Unreviewed) view filters — pull up just your worklist (unreviewed), audit everything excluded, and so on. After an initial pass through your library, "unreviewed" doubles as "new since I last looked." Per-title and bulk.
• Browse ⇄ De-duplicated review toggle — switch between the classic per-category tree and the de-dup list; both edit the same exclusion list, so switching is lossless, and your choice is remembered.
• Title-level series exclusion — Dispatcharr gives the same show a distinct ID per category, so excluding it in one place can leave copies elsewhere. The sync applies your exclusions to a title's whole group of duplicate copies, not just the IDs you ticked, so a copy that appears later under a fresh ID is skipped without any action from you. This is based on the fact that Dispatcharr uses the exact same title for the series for every merged copy of it. Titles that Dispatcharr hasn't merged and have different names will appear as different items in the review interface.
• Only sync movies and series you have reviewed — normally anything you have not excluded gets synced, which is fine until a provider adds content in bulk. With this option on, a title is written only once you have reviewed it or excluded it, so new arrivals wait in the de-dup view instead of landing in your library. If a title already on disk reappears with a different ID it is automatically marked as reviewed and continues to sync. For series, an existing record of their episodes counts as recognition too, so a rename is fine as long as the ID is stable. Turning on metadata IDs in folder names makes this recognition considerably more reliable. If the reviewed-list setting is ever unreadable the option disables itself for that run and says so in the log, rather than treating every title as unreviewed and holding back your whole library.
• Sync robustness for duplicates — cross-listed series collapse to one folder instead of writing duplicate per-episode files, and series whose episode list returns empty under load are retried so a batch of new titles lands in one sync.

Surviving a provider renumbering its catalog

Every exclusion and reviewed mark is stored against the provider's stream ID, and some providers periodically reissue those in bulk. When that happens, thousands of decisions silently detach at once: long-settled titles reappear as unreviewed, files pointing at dead IDs get cleaned up as orphans, and nothing in any log explains it. This fork treats that as a first-class failure rather than bad luck.

• Decisions re-point themselves — the plugin records each movie's TMDB ID alongside the decision, so a title returning under a new ID is recognised and your choice moves with it, during the ordinary sync. There is deliberately no button and no prompt: the defining property of this failure is that nothing tells you it happened, so anything needing you to notice has already failed. It protects decisions from the moment it is installed onward, needs the provider to supply a TMDB ID, and does not cover series, whose listing carries none.
• The sync will not delete a folder it just wrote — exclusions are stored per ID but enforced by folder name, so two catalog entries that resolve to one folder name used to make the sync write a title and then immediately delete it, on every run, leaving you with neither copy.
• It keeps its own records, with no setup — the store sizes in a log that does not rotate, a dated listing of which provider ID was which title, a rollback copy taken immediately before every configuration write, and a scheduled backup of the whole configuration. Point Backup and records folder at a different drive and the backup, the listings and the store-size log move there, which is what turns them from same-disk copies into real ones. The rollback stays beside the configuration on purpose: it is taken on every write, so it has to be somewhere always available.
• Recovery scripts for what the plugin cannot reach — damage that predates installation, series identity, and repairing a configuration after the fact. See Diagnostics and recovery scripts.

The in-plugin half exists because the alternative did not hold up in practice: tooling that needs Docker, a cron entry and prior knowledge protects the people who least need protecting. A snapshot is only useful if it was taken before the event, which means it cannot depend on anyone remembering to take one.

Related projects

This plugin is one of four small projects that together run a Dispatcharr-backed VOD library in Emby. Each is useful on its own, but they were built to fit together, and one dependency is worth stating plainly before you rely on this one.

• dispatcharr_vod_concurrency_fix — coalesces the near-simultaneous range requests some clients (notably Emby) make when playing MKV VOD files, so they share one provider slot instead of failing over to a different file and corrupting playback.
• dispatcharr_vod_preferences — control over which VOD stream Dispatcharr serves through its proxy for a given title, which clients cannot otherwise reach. Composable options to prefer better video or audio quality, or to remember a specific stream or provider.
• dispatcharr_vod_episode_sweep — learns which VOD series a client syncs, then once a day refreshes all of each watched show's provider and category relations, so new episodes stop going missing. Scoped to only the series you sync; runs in the background.
• emby-xtream-dedupe (this repo) — generates .strm files for Emby from a Dispatcharr catalog, de-duplicating titles cross-listed across categories and holding new arrivals for review before they reach your library.

A .strm generator alone will not keep your episodes up to date

A .strm generator writes files for the episodes Dispatcharr already knows about. It does not — and cannot — make Dispatcharr go and look for new ones. Dispatcharr refreshes a series' episode data lazily, and nothing in a normal sync forces that refresh: a healthy, settled library makes no episode-detail calls at all, precisely because nothing has changed from its point of view.

So without a server-side sweep, new episodes of shows you already have can simply never appear. The library looks healthy, the syncs report success, and the missing episodes are invisible because nothing ever asked for them.

dispatcharr_vod_episode_sweep exists to close exactly that gap, and is designed to run alongside a .strm generator like this one. It learns which series a client actually syncs, so it stays scoped to your library rather than hammering the whole catalog, and refreshes every provider and category relation of each show once a day. Any equivalent server-side refresh will do — the requirement is that something periodically makes Dispatcharr re-check its providers for new episodes. If you run this plugin without one, treat missing episodes as expected rather than as a bug here.

Features

Live TV & EPG

Full Live TV integration with Emby's native TV guide.

• M3U playlist generation with channel metadata, logos, and EPG channel IDs
• XMLTV electronic program guide with configurable fetch window (1-14 days)
• Category-based filtering — select which channel groups to include
• Stream format selection — MPEG-TS or HLS (M3U8)
• Adult content filtering — opt-in toggle for adult-flagged channels
• DVB subtitle declaration — optional toggle that surfaces DVB subtitle tracks for live TV without re-enabling stream probing (see ADR-009)
• Automatic caching — M3U (15 min) and EPG (30 min) with thread-safe invalidation

VOD Movie Library

Sync on-demand movies as STRM files that Emby treats as a native movie library.

• STRM file generation — one file per movie, Emby handles metadata and artwork
• Folder organization modes:
  • Single folder — all movies in one Movies/ directory
  • Multiple folders — auto-organized by provider category name
  • Custom mapping — define your own folders and assign categories to each
• TMDB metadata matching — appends [tmdbid=123] to folder names for instant Emby identification
• TMDB fallback lookup — queries Emby's metadata providers when the Xtream source lacks a TMDB ID
• Category selection — pick specific VOD categories to sync, or sync all
• Per-title selection — expand any category (Single Folder mode) to tick individual movies. Unticked titles stop syncing and their folder is removed on the next sync, whether or not orphan cleanup is on
• Danger zone — one-click delete of all synced movie content

TV Series Library

Full series support with proper season/episode structure.

• Season/Episode STRM files — Show Name/Season 01/Show Name - S01E01 - Episode Title.strm
• Series detail fetching — pulls episode lists per series from the Xtream API
• TVDb / TMDB ID folder naming — Show Name [tvdbid=81189] for reliable metadata matching
• Manual ID overrides — force a specific TVDb ID for shows that don't auto-match
• Metadata fallback lookup — searches Emby's providers when no ID is available
• Same folder modes as movies — single, multiple, or custom category mapping
• Per-title selection — same as movies, at whole-series granularity rather than per-episode

Smart Sync Engine

Efficient sync that doesn't re-download what you already have.

• Smart skip — skips writing STRM files that already exist on disk
• Configurable parallelism — 1-10 concurrent operations (default 3)
• Orphan cleanup — automatically removes STRM files for content no longer in the source
• Cross-listing deduplication — movies/series appearing in multiple categories are synced once
• Content name cleaning — strips provider prefix tags (e.g. |UK|, |FR|) and custom terms from titles
• Real-time progress — Phase, Total, Completed, Skipped, Failed counters polled every 500ms

Dispatcharr Integration

Optional integration with Dispatcharr for IPTV stream management.

• Stream proxy routing — routes Live TV through Dispatcharr's proxy for connection management
• Pre-populated media info — fetches codec, resolution, and bitrate from Dispatcharr's stream stats (requires Streamflow configured in Dispatcharr to generate per-channel metadata)
• FFprobe bypass — skips Emby's stream analysis when stats are available (faster channel switching)
• JWT authentication — automatic token refresh with retry and exponential backoff
• Graceful fallback — reverts to direct Xtream URLs if Dispatcharr is unavailable

Built-in Dashboard

A configuration UI embedded in Emby's plugin settings with five tabs.

• Dashboard — last sync status, sync history (last 10), library stats, live progress bar
• Settings — server connection, sync tuning, name cleaning, metadata matching
• Movies — enable/disable, folder mode, category selection with search, sync button
• Series — same layout as movies with series-specific options
• Live TV — stream format, EPG settings, catch-up, Dispatcharr, category filtering

---

Installation

Step 1: Download the Plugin

Download the DLL matching your Emby Server version from the latest release:

Your Emby Server	Download	Install as
4.9.x	Emby.Xtream.Plugin.dll	Emby.Xtream.Plugin.dll
4.10.0.17 and later	Emby.Xtream.Plugin-4.10.dll	rename to Emby.Xtream.Plugin.dll

Emby 4.10 left beta and reached general release in September 2026, so most installs now want the second one. The two builds target different Emby SDKs and are not interchangeable — check your version under Dashboard → Help → About if you are unsure. Install one, not both.

⚠️ Rename the 4.10 download before installing it

The two builds cannot share a filename in the same release, so the 4.10 one is published with a -4.10 suffix. Emby names each plugin's settings file after the DLL, so a file left as Emby.Xtream.Plugin-4.10.dll gets its own separate settings file — the plugin loads, the settings page opens, and everything you had configured appears blank, because your real settings are still sitting in the file belonging to the other name.

Nothing is lost when this happens and nothing warns you either. Rename the file to Emby.Xtream.Plugin.dll before copying it in and your existing configuration is picked up normally.

The built-in update check reads your Emby version and downloads the build matching it, writing over the file already in place. So installing an update never moves you onto the wrong build. Changing Emby's own major version is still worth doing by hand, though: the plugin has to load before it can check for anything, and a build meant for the other Emby may not get that far.

Only the single DLL file is needed — no other dependencies.

Build from source (alternative)

Requires .NET SDK 6.0+:

git clone https://github.com/andyj682/emby-xtream-dedupe.git
cd emby-xtream-dedupe
bash build-dedupe.sh

That builds both Emby targets, runs both test configurations, and checks the delete-site guard — the same things CI does. The DLLs land at:

Target	Path
Emby 4.9.x	Emby.Xtream.Plugin/out/Emby.Xtream.Plugin.dll
Emby 4.10.0.17+	out_4_10/Emby.Xtream.Plugin.dll

Both are already named Emby.Xtream.Plugin.dll, which is what Emby needs — only the published release asset carries the -4.10 suffix, and that one has to be renamed on install.

Expect the 4.10 test run to report more tests than the 4.9 one: a handful of them cover methods that only exist in that build.

To build just the 4.9 target, cd Emby.Xtream.Plugin && bash build.sh still works.

Step 2: Install the Plugin

Copy the DLL to your Emby Server's plugins directory and restart.

Docker (most common):

docker cp Emby.Xtream.Plugin.dll emby:/config/plugins/
docker restart emby

Bare metal (Linux):

cp Emby.Xtream.Plugin.dll /var/lib/emby/plugins/
systemctl restart emby-server

Bare metal (macOS/Windows): Copy Emby.Xtream.Plugin.dll to your Emby data directory under plugins/, then restart Emby Server.

Step 3: Configure the Plugin

1. Open Emby's web UI
2. Go to Settings > Plugins > Xtream Tuner
3. Enter your Xtream server details:
   • Server URL — e.g. http://your-provider:port
   • Username and Password
4. Click Test Connection to verify
5. Click Save

Step 4: Set Up Live TV

1. Switch to the Live TV tab
2. Choose your Stream Format (MPEG-TS recommended)
3. Click Refresh Categories to load channel groups
4. Select the categories you want
5. Configure EPG settings (days to fetch, cache duration)
6. Click Save
7. Go to Emby Settings > Live TV and add a new tuner:
   • Type: Xtream Tuner
   • It will auto-discover the plugin's M3U and EPG endpoints

Step 5: Set Up Movies (Optional)

1. Switch to the Movies tab
2. Check Enable VOD Movies
3. Click Refresh Categories to load VOD categories
4. Choose a Folder Organization mode:
   • Single Folder — select categories, all movies go to Movies/
   • Multiple Folders — one folder per category, auto-named
   • Custom — click "Add Folder", name it, assign categories
5. Click Sync Movies Now
6. In Emby, add a new Movies library pointing to the STRM output path (default: /config/xtream/Movies)

Step 6: Set Up Series (Optional)

1. Switch to the Series tab
2. Check Enable Series / TV Shows
3. Same workflow as Movies — refresh categories, select, choose folder mode
4. Click Sync Series Now
5. In Emby, add a new TV Shows library pointing to /config/xtream/Shows

Step 7: Dispatcharr Integration (Optional)

If you use Dispatcharr for stream management:

1. Go to Live TV tab > Dispatcharr section
2. Check Enable Dispatcharr
3. Enter Dispatcharr URL, username, and password
4. Click Test Dispatcharr to verify
5. Save — Live TV streams will now route through Dispatcharr's proxy

Note: For the plugin to receive codec/resolution metadata and skip FFprobe, Streamflow must be enabled and configured in Dispatcharr. Without Streamflow generating per-channel stream stats, the plugin falls back to standard stream handling.

Updating the Plugin

Download the latest DLL from Releases, replace the file in your plugins directory, and restart Emby. Your configuration is preserved across updates.

---

Configuration Reference

Setting	Default	Description
Stream Format	MPEG-TS	Live TV container format (ts or m3u8)
Enable DVB Subtitles	Off	Declare DVB subtitle tracks on every live channel so they appear without stream probing (non-DVB sources will show empty subtitle options)
EPG Cache	30 min	How long to cache EPG data (5-1440 min)
EPG Days	2	Days of guide data to fetch (1-14)
M3U Cache	15 min	How long to cache channel playlists (1-1440 min)
STRM Library Path	/config/xtream	Where STRM files are written
Smart Skip	On	Skip existing STRM files during sync
Sync Parallelism	3	Concurrent operations during sync (1-10)
Cleanup Orphans	Off	Remove STRM files not in source
Only sync movies and series you have reviewed	Off	Hold titles that are neither reviewed nor excluded instead of writing them. Titles already on disk are exempt and marked reviewed automatically. Works best with metadata IDs in folder names
Refresh Emby libraries after a sync	On	Tell Emby a library changed once a sync has actually added or removed files
TMDB Folder Naming	Off	Append [tmdbid=X] to movie/series folders
Fallback Lookup	Off	Query Emby's metadata providers for missing IDs
Name Cleaning	Off	Strip prefix tags and custom terms from titles
Backup and records folder	(blank)	Where configuration backups, catalog snapshots and the store-size history are kept (RecordsPath). Blank means xtream-backups beside the plugin's configuration. See below
Configuration backups to keep	10	Daily copies of your exclusions, reviewed marks and settings, taken by a scheduled task (ConfigBackupCount; 0 disables)
Configuration rollback copies to keep	10	Copies taken immediately before the plugin writes its configuration, kept beside it (ConfigRollbackCount; 0 disables). See below
Catalog snapshots to keep	10	Dated listings of which provider ID was which title, written during each sync (CatalogueSnapshotCount; 0 disables)
Wanted list folder	(blank)	Where to write wanted-set.json, the list of movies you sync, for a Dispatcharr plugin to read (WantedSetPath). Blank turns it off. Dispatcharr must be able to read the folder; see below

---

Protecting your configuration

Your exclusions and reviewed marks are the expensive part of this plugin's state. On a curated library that is tens of thousands of individual decisions, and nothing else on your system knows what they were — a media scan cannot rebuild them.

What the plugin does for you, with no setup at all:

• Store sizes in every sync log, so a sudden drop is visible rather than something you notice weeks later when the review queue looks wrong — and appended to counts.log as well, because Emby's log rotates and the value of those counts is the trend.
• The full list of deleted files beside Emby's logs whenever a cleanup removes more than the log line samples.
• A rollback copy of the configuration into an xtream-rollback folder next to it, taken immediately before the plugin writes.
• A scheduled backup of the configuration, and a dated catalog snapshot written during each sync — both under the Backup and records folder, which defaults to xtream-backups beside the configuration.

Putting one back. Under Restore a Saved Configuration on the Settings tab, Show saved copies lists every rollback and backup with the date it was taken and what restoring it would change — "nothing", or something like "2 settings, +8,554 movie exclusions" — so you can tell them apart without having to remember what was in them. Restoring replaces your whole configuration with that copy — every setting, not only the exclusions and reviewed marks — and takes a rollback copy of the current state first, so a restore can itself be undone. You start it; the plugin never restores anything on its own.

A copy whose decision stores cannot be read is listed but refused, because restoring it would replace a readable store with an unreadable one. Restoring is also refused while a sync is running, since the sync would write its own results over it moments later.

What it does not do. By default all of that lives on the same disk as the file it protects. It covers a bad write; it does not survive losing the volume. And if you point the backup folder at another drive, note that the pointer lives in the configuration itself — so after losing the configuration entirely, the plugin looks in the default location and will not find copies kept elsewhere. Put them back by hand first, then restore.

So point the backup folder at a different drive if you have one — that single setting upgrades the backups and snapshots from same-disk to real ones. For anything off-machine, the whole configuration is still a single XML file:

cp /path/to/emby/config/plugins/configurations/Emby.Xtream.Plugin.xml ~/backups/$(date +%F).xml

⚠️ That file — and every rollback copy — contains your provider username and password in plain text. Keep the copies local, and never commit one to version control or attach one to a bug report.

---

Diagnostics and recovery scripts

scripts/ holds standalone Python tools for answering "what does my live data actually say" without a rebuild or a deploy. They read the plugin's config XML and the provider's list endpoints only — never get_series_info, which trips Dispatcharr's gated episode refresh — so they are safe to run against a working setup. Nothing writes to the live config.

Script	What it answers
catalogue-snapshot.py	Records which provider ID was which title, today, plus a census of how many stored IDs no longer exist. Run it periodically; the others consume its output
audit-strm-links.py	Which .strm files point at a stream ID the provider no longer has, split into "the title came back under a new ID" and "the content is gone"
analyse-new-arrivals.py	Given two snapshots, splits a batch of new arrivals into already-excluded / already-reviewed / already-on-disk / genuinely new — i.e. how much of a review queue is actually new work
repair-id-churn.py	Re-points exclusions and reviewed marks after a provider reassigns IDs. Dry run by default; --write emits a candidate config for you to diff and install yourself
check-repair-safety.py	Run before installing a repair-id-churn.py candidate: would any proposed exclusion land on a title that is currently on disk? Checks both the stream ID and the folder name, because exclusion is stored per ID but enforced per name
config-counts-canary.py	How many exclusions and reviewed marks does a config actually contain? Works on the live config, a test rig's, or a repair candidate. Distinguishes "empty" from "unreadable", which look identical in a count and mean opposite things
find-crosslisted-series.py	Finds a cross-listed series, reports titles that are only partly excluded, and can emit a minimal test config
xtream_catalogue.py	Shared helpers (not run directly)

config-counts-canary.py is worth running on a schedule with --log: the four counts only mean something as a trend, and a silent drop is the failure you most want to catch early. It exits non-zero when a store is unreadable, so it works as a cron alarm.

They assume Docker and a throwaway python:3-alpine container, e.g.:

docker run --rm --network container:emby --memory=512m \
  -v /etc/localtime:/etc/localtime:ro \
  -v /path/to/emby/config:/cfg:ro \
  -v "$PWD/scripts":/scripts:ro \
  -v "$HOME/xtream-snapshots":/out \
  python:3-alpine python3 /scripts/catalogue-snapshot.py --out /out

--network container:emby matters if your Xtream base URL is a hostname on a Docker network — the default bridge cannot resolve it, and the failure looks like a DNS error rather than a credentials problem. Note also that snapshots contain titles, and the config XML they read contains your provider credentials in plaintext (see below), so keep both local.

-v /etc/localtime:/etc/localtime:ro matters more than it looks. These scripts stamp dated filenames and log lines with the container's local time, and a bare python:3-alpine has no timezone set, so it runs in UTC — while Emby, and therefore the plugin's own records, use whatever timezone that container was given. Without the mount the two disagree by your UTC offset: a snapshot taken in the evening can land under tomorrow's date, and a counts log interleaved with the plugin's own would run backwards. Mounting the host clock makes every record agree, and it needs no tzdata package.

---

Logging wanted movies for Dispatcharr

Providers describe TV episodes well and movies poorly, so a Dispatcharr plugin that picks the best stream for each title, or fills in missing stream details, can do it for episodes and not for films. Filling that gap means probing movies, and that needs a list of the ones worth probing. This plugin is the only thing that knows which movies you sync, so it can write that list down.

Set Wanted list folder under Log Wanted Movies for Dispatcharr in Settings, and after every movie sync the plugin writes wanted-set.json there:

{
  "schema": 1,
  "generated_at": "2026-09-16T04:12:33Z",
  "generator": "emby-strm",
  "count": 2,
  "tmdb_ids": [603, 27205],
  "unidentified": [{ "stream_id": 419883, "name": "Example Film (2019)" }]
}

• count covers tmdb_ids only, so a reader can tell a complete file from a partial one.
• unidentified lists wanted movies the provider gave no TMDB ID for, by stream ID and the provider's own name for the title, exactly as sent. Stream IDs change when a provider renumbers its catalog; names usually do not.
• resolved_tmdb_id can also appear on those entries, but only with TMDB Folder Naming and Fallback Lookup both on, so most setups will not see it.
• generated_at is UTC.

The file is rewritten in full on each sync and holds nothing you would lose by deleting it. If any VOD category fails to answer, the plugin skips the write and leaves the previous file in place, because a list missing a whole category looks exactly like a shorter list you chose.

Important

This does nothing unless Dispatcharr can read the folder, and a wrong path looks exactly like a working one. The simplest arrangement is a folder beside Emby's config, mounted into Emby read-write and into Dispatcharr read-only:

# Emby
- /path/to/emby/exchange:/exchange
# Dispatcharr
- /path/to/emby/exchange:/xtream-exchange:ro

Set the folder to /exchange here, and point the reading plugin at /xtream-exchange. Confirm from Dispatcharr's side with docker exec dispatcharr ls /xtream-exchange; Emby's log reporting a successful write only proves the file went somewhere.

The list contains real titles. It holds no credentials, but treat it like a catalog snapshot and keep it on your own host. If Only sync movies and series you have reviewed is off, the list is your whole included catalog rather than the titles you chose, and the plugin notes that in the log. grep -i "wanted set" in Emby's log shows each write.

How a particular Dispatcharr plugin uses the file is documented with that plugin. The format and the reasoning behind it are in ADR-F008.

---

Security Note: Credentials in STRM Files and M3U Output

The Xtream Codes protocol requires the username and password to appear directly in every stream URL — there is no token or session form. As a result, your Xtream credentials end up embedded in plaintext in two places:

• The .strm files written under your STRM Library Path (default /config/xtream)
• The M3U playlist served by the plugin's tuner endpoint

This is a property of the protocol, not a plugin defect — the same is true of any Xtream-based tooling. Anything that can read those files (other containers sharing the volume, backup tools, indexers, users with shell access on the host) can harvest the credentials.

How to keep credentials off disk:

• Use Dispatcharr. When Dispatcharr is enabled, Live TV and multi-provider VOD URLs route through Dispatcharr's credential-free proxy URLs (/proxy/ts/stream/<uuid>, /proxy/vod/movie/<uuid>) instead of the raw Xtream form. Single-provider VOD movies and series episodes still fall back to the Xtream URL today.
• Restrict filesystem permissions. Make sure the STRM library path is not readable by other users or containers on the host. For Docker, avoid mounting /config/xtream into unrelated services.
• Use a dedicated provider account. If your Xtream provider supports multiple credential sets per subscription, use a separate one for Emby so it can be rotated independently.

If credential exposure on disk is unacceptable for your environment, Dispatcharr in front of the plugin is currently the only complete answer. Improving credential-free routing for single-provider VOD and series episodes is tracked as a future enhancement.

---

License

MIT — see LICENSE.

This is a fork, so the copyright notice carries two lines: one for firestaerter3/emby-xtream, whose code this is built on and which declares MIT, and one for the additions made here. Both are under the same terms.