Worker CLI Reference¶

The mediafusion-worker binary supports a one-shot CLI mode that lets you trigger any registered job directly from the command line without starting the full scheduler. This is useful for manual scraping runs, debugging, and CI pipelines.

Usage¶

mediafusion-worker [--list-jobs] [--run-job <queue>] [--args <json>]

With no arguments the worker starts normally and runs the full scheduler.

Flags¶

`--list-jobs`¶

Prints every registered queue name and exits immediately. Use this to discover available job names.

BinaryDockerDocker ComposeMakefile (dev)

./mediafusion-worker --list-jobs

docker run --rm --env-file .env mhdzumair/mediafusion:6.0.0 \
  mediafusion-worker --list-jobs

docker compose exec mediafusion-worker \
  /mediafusion-worker --list-jobs

make worker-list-jobs

`--run-job <queue>`¶

Runs a single job inline and exits. The worker connects to the database and Redis, runs the job to completion, then exits with code 0 on success or 1 on failure.

BinaryDockerDocker ComposeMakefile (dev)

./mediafusion-worker --run-job spider_sport_video

docker run --rm --env-file .env mhdzumair/mediafusion:6.0.0 \
  mediafusion-worker --run-job spider_sport_video

docker compose run --rm mediafusion-worker \
  /mediafusion-worker --run-job spider_sport_video

make worker-run-job JOB=spider_sport_video

`--args <json>`¶

Passes a JSON payload to the job. Required for jobs that take parameters (see Jobs that accept arguments below). Ignored by jobs that don't use it.

./mediafusion-worker --run-job dmm_hashlist --args '{"full": true}'

With the Makefile:

make worker-run-job JOB=dmm_hashlist JOB_ARGS='{"full": true}'

All registered jobs¶

Run --list-jobs to get the live list. The table below describes each queue.

Background scrapers¶

Queue	Description
`spider_tamilmv`	Scrape TamilMV for Tamil content
`spider_tamil_blasters`	Scrape Tamil Blasters
`spider_formula_ext`	Scrape Formula Racing (ext source)
`spider_motogp_ext`	Scrape MotoGP (ext source)
`spider_wwe_ext`	Scrape WWE (ext source)
`spider_ufc_ext`	Scrape UFC (ext source)
`spider_movies_tv_ext`	Scrape Movies/TV (ext source)
`spider_sport_video`	Scrape sport-video.org.ua
`spider_registry_crawl`	Run all public indexer spiders (1337x, TPB, YTS, Nyaa, AnimeTosho, SubsPlease, AnimePahe, BT4G, EZTV, Rutor, LimeTorrents, BT52, UIndex)
`spider_eztv_rss`	Scrape EZTV RSS feed
`acestream_bg`	Scrape AceStream channels
`youtube_bg`	Scrape YouTube channels
`telegram_bg`	Scrape configured Telegram channels

TRAWL-dependent spiders

spider_movies_tv_ext, spider_formula_ext, spider_motogp_ext, spider_wwe_ext, spider_ufc_ext, spider_sport_video, and CF entries in spider_registry_crawl require TRAWL_URL to be set. See TRAWL setup.

Feed scrapers¶

Queue	Description
`prowlarr_feed`	Fetch new releases from Prowlarr indexers
`jackett_feed`	Fetch new releases from Jackett indexers
`rss_feed`	Process all configured RSS feeds
`dmm_hashlist`	Sync DebridMediaManager hashlist from GitHub

Imports¶

Queue	Description
`m3u_import`	Import/sync a single M3U IPTV source
`xtream_import`	Import/sync a single Xtream Codes source
`imdb_dataset_import`	Import IMDb non-commercial datasets into the DB

Search & maintenance¶

Queue	Description
`background_search`	Process the background re-scrape queue
`discover_prewarm`	Pre-warm Discover catalog caches
`integration_syncs`	Sync Trakt / Simkl watchlists
`update_seeders`	Update seeder counts for tracked torrents
`update_tv_posters`	Refresh TV show poster images
`validate_tv`	Validate and deactivate dead TV stream URLs
`backfill_stream_metadata`	Re-parse torrent/usenet `stream.name` with PTT; fill missing resolution/quality/language/HDR/audio links
`cleanup`	Remove expired scraper task records and cache entries

Jobs that accept arguments¶

Most jobs ignore --args. The following jobs use it:

`dmm_hashlist`¶

Field	Type	Default	Description
`full`	`bool`	`false`	Walk the full commit history instead of only new commits
`reset_checkpoints`	`bool`	`false`	Clear saved Redis checkpoints before running

# Incremental sync (normal)
./mediafusion-worker --run-job dmm_hashlist

# Full re-sync from the beginning
./mediafusion-worker --run-job dmm_hashlist --args '{"full": true}'

# Reset checkpoints then sync
./mediafusion-worker --run-job dmm_hashlist --args '{"full": true, "reset_checkpoints": true}'

`m3u_import`¶

Field	Type	Required	Description
`iptv_source_id`	`int`	yes	Database ID of the `iptv_source` row to import

./mediafusion-worker --run-job m3u_import --args '{"iptv_source_id": 42}'

`xtream_import`¶

Field	Type	Required	Description
`iptv_source_id`	`int`	yes	Database ID of the `iptv_source` row to import

./mediafusion-worker --run-job xtream_import --args '{"iptv_source_id": 7}'

`imdb_dataset_import`¶

Field	Type	Default	Description
`datasets`	`string[]`	all datasets	Subset of dataset keys to process

# Import all datasets
./mediafusion-worker --run-job imdb_dataset_import

# Import only specific datasets
./mediafusion-worker --run-job imdb_dataset_import \
  --args '{"datasets": ["title.basics", "title.akas"]}'

`backfill_stream_metadata`¶

Re-parses existing torrent and usenet stream release names through the PTT parser (parser::parse_title, with translate_languages: true — same as Python PTT.parse_title(name, True)) and writes any missing metadata:

stream columns: resolution, codec, quality, release_group (and flags when only_missing is false)
Link tables: stream_language_link, stream_hdr_link, stream_audio_link, stream_channel_link

By default the job runs continuously in one process until no streams match the criteria (suitable for CLI one-shot runs). Set "continuous": false to process a single page and enqueue the next page on the worker job queue instead.

Field	Type	Default	Description
`after_id`	`int`	`0`	Resume after this stream id (keyset pagination)
`batch_size`	`int`	`500`	Streams per batch (max 5000). Alias: `page_size`
`only_missing`	`bool`	`true`	Only streams missing columns or link-table rows
`stream_types`	`string[]`	`["TORRENT","USENET"]`	Stream types to process
`continuous`	`bool`	`true`	Loop batches in-process until done
`max_batches`	`int`	—	Optional cap (for test runs). Alias: `max_pages`

# Full backfill (runs until complete — may take hours on large DBs)
./mediafusion-worker --run-job backfill_stream_metadata

# Larger batches, same behavior
./mediafusion-worker --run-job backfill_stream_metadata \
  --args '{"batch_size": 2000}'

# Resume after a specific stream id (e.g. after stopping a run)
./mediafusion-worker --run-job backfill_stream_metadata \
  --args '{"after_id": 1234567, "batch_size": 2000}'

# Test run: first 3 batches only
./mediafusion-worker --run-job backfill_stream_metadata \
  --args '{"batch_size": 500, "max_batches": 3}'

# Re-parse everything (overwrite columns from PTT; links are additive)
./mediafusion-worker --run-job backfill_stream_metadata \
  --args '{"only_missing": false, "batch_size": 1000}'

Makefile shortcut:

make worker-backfill-stream-metadata
make worker-backfill-stream-metadata BATCH_SIZE=2000
make worker-backfill-stream-metadata AFTER_ID=1234567 BATCH_SIZE=2000

After backfill, Stremio stream descriptions should show language lines (🌐 English, etc.) and language filtering in the addon will work for streams that PTT can parse.

Exit codes¶

Code	Meaning
`0`	Job completed successfully
`1`	Job failed or unknown `--run-job` queue name

Examples¶

# Manually trigger the sport video scraper
./mediafusion-worker --run-job spider_sport_video

# Run all public indexer spiders
./mediafusion-worker --run-job spider_registry_crawl

# Force a full DMM hashlist re-sync
./mediafusion-worker --run-job dmm_hashlist --args '{"full": true, "reset_checkpoints": true}'

# Process the background re-scrape queue once
./mediafusion-worker --run-job background_search

# Sync Trakt/Simkl integrations
./mediafusion-worker --run-job integration_syncs

# Backfill PTT metadata on existing torrent/usenet streams
./mediafusion-worker --run-job backfill_stream_metadata

# List all available jobs
./mediafusion-worker --list-jobs