Content Sources¶

Configure where MediaFusion fetches stream data. These are the scrapers, indexers, and feed integrations that populate your catalogs.

Built-in stream sources¶

Variable	Default	Description
`IS_SCRAP_FROM_TORRENTIO`	`true`	Fetch streams from Torrentio
`IS_SCRAP_FROM_MEDIAFUSION`	`true`	Fetch from the community MediaFusion index
`IS_SCRAP_FROM_ZILEAN`	`true`	Fetch cached content from Zilean DMM

Prowlarr integration¶

Prowlarr aggregates torrent indexers and is the recommended way to get comprehensive torrent results.

Variable	Default	Description
`PROWLARR_URL`	`http://prowlarr:9696`	Prowlarr base URL
`PROWLARR_API_KEY`	`None`	Prowlarr API key (generate in Prowlarr → Settings → General)
`PROWLARR_IMMEDIATE_MAX_PROCESS`	`30`	Max simultaneous Prowlarr searches on live requests
`PROWLARR_IMMEDIATE_MAX_PROCESS_TIME`	`30`	Timeout in seconds for live Prowlarr searches
`PROWLARR_SEARCH_INTERVAL_HOUR`	`24`	How often to re-search Prowlarr for cached titles (hours)
`PROWLARR_LIVE_TITLE_SEARCH`	`true`	Search Prowlarr in real time when a stream is requested

See Prowlarr Integration for setup instructions.

Torznab endpoints¶

Add custom Torznab-compatible indexers beyond Prowlarr:

IS_SCRAP_FROM_TORZNAB=true
TORZNAB_ENDPOINTS='[
  {
    "name": "My Indexer",
    "url": "https://indexer.example/api?apikey=xxx",
    "enabled": true,
    "priority": 1,
    "categories": [2000, 5000]
  }
]'

Background scrapers¶

These scrapers run on a schedule to keep catalogs fresh. Enable the ones relevant to your use case:

Variable	Default	Description
`IS_SCRAP_FROM_ACESTREAM_BACKGROUND`	`true`	Scrape AceStream channels in the background
`ACESTREAM_BACKGROUND_SEARCH_API_KEY`	`None`	API key for AceStream search
`IS_SCRAP_FROM_YOUTUBE_BACKGROUND`	`false`	Scrape YouTube content
`YOUTUBE_API_KEY`	`None`	Required when YouTube scraping is enabled
`IS_SCRAP_FROM_TELEGRAM_BACKGROUND`	`false`	Scrape Telegram channels

See Telegram Integration for bot setup, per-user sessions, channel configuration, and scrape depth options.

Telegram channel scraping¶

Per-user Telegram sessions scrape channels/groups the user has joined. Requires TELEGRAM_API_ID, TELEGRAM_API_HASH, and bot configuration.

Variable	Default	Description
`TELEGRAM_API_ID`	—	Telegram user API ID (from my.telegram.org)
`TELEGRAM_API_HASH`	—	Telegram user API hash
`MIN_SCRAPING_VIDEO_SIZE`	`26214400`	Minimum video size in bytes (25 MB)
`TELEGRAM_BACKGROUND_SCRAPER_CRONTAB`	(built-in)	Cron for the background scraper
`DISABLE_TELEGRAM_BACKGROUND_SCRAPER`	`false`	Disable scheduled Telegram scraping

Manual scrapes (web UI or /scrape in the bot) default to 25 messages per channel; users can choose a custom count or full history. See the Telegram integration guide for details.

Live search¶

Variable	Default	Description
`LIVE_SEARCH_STREAMS`	`true`	Fan out to N scrapers in parallel when a stream is requested. Adds 1–5 s latency but surfaces fresher results.

Disable this to reduce latency at the cost of fewer real-time results.

TRAWL (browser-based scraping)¶

TRAWL is the browser pool used by Rust v6 workers to bypass Cloudflare and JS bot challenges. It exposes a FlareSolverr-compatible /v1 API.

Set TRAWL_URL to the TRAWL base URL (e.g. http://trawl:8191 in Docker Compose). When unset, CF-protected scrapers are skipped or degraded:

Spider / source	Behaviour without TRAWL
ext.to (`spider_*_ext`)	Listing pages may work; magnet/torrent AJAX downloads fail
sport-video (`spider_sport_video`)	Job exits immediately with a warning
Public indexers (1337x, TPB, …)	CF indexers skipped during `spider_registry_crawl`
Formula feeds (Reddit RSS)	Direct HTTP only; no browser fallback

Variable	Default	Description
`TRAWL_URL`	—	TRAWL base URL (FlareSolverr-compatible API at `/v1`).
`BYPARR_URL`	—	Deprecated alias for `TRAWL_URL`.

In Docker Compose, TRAWL shares the MediaFusion Redis instance on database 1 (redis://redis:6379/1). MediaFusion uses database 0. Key namespaces do not overlap.

Python v5 only (deprecated)¶

These variables applied to the removed Scrapling/Browserless stack in python-deprecated/:

Variable	Default	Description
`SCRAPLING_CDP_URL`	`None`	WebSocket CDP endpoint for deprecated Scrapling workers
`SCRAPLING_FETCHER_MODE`	`stealthy`	`stealthy` or `dynamic`
`SCRAPLING_HEADLESS`	`true`	Run browser headlessly
`SCRAPLING_SOLVE_CLOUDFLARE`	`true`	Attempt Cloudflare bypass
`SCRAPLING_REAL_CHROME`	`false`	Use a real Chrome binary instead of Playwright
`SCRAPLING_PROXY_URL`	`None`	Proxy for scrapling requests
`PUBLIC_INDEXERS_LIVE_SEARCH_ENABLE_CLOUDFLARE_SOLVER`	`false`	Enable Cloudflare solver for live searches (Python v5)

Cloudflare solver

Rust v6 workers use TRAWL_URL for background jobs only — live /stream/ requests never invoke TRAWL inline. Leave PUBLIC_INDEXERS_LIVE_SEARCH_ENABLE_CLOUDFLARE_SOLVER=false unless you still run the deprecated Python server.

Proxy settings¶

Variable	Default	Description
`REQUESTS_PROXY_URL`	`None`	Route all outbound HTTP (scrapers + debrid API calls) through this proxy
`REQUESTS_PROXY_EXCLUDE_DEBRID_PROVIDERS`	`[]`	Comma-separated (or JSON array) debrid provider IDs that bypass the proxy and connect directly. Ignored when `REQUESTS_PROXY_INCLUDE_DEBRID_PROVIDERS` is set. E.g. `realdebrid,torbox`. Valid IDs: `realdebrid`, `seedr`, `debridlink`, `alldebrid`, `offcloud`, `pikpak`, `torbox`, `premiumize`, `stremthru`, `easydebrid`, `debrider`.
`REQUESTS_PROXY_INCLUDE_DEBRID_PROVIDERS`	`[]`	When non-empty, only these debrid provider IDs are routed through the proxy; all others connect directly. Takes precedence over the exclude list. Same format and valid IDs as above.
`REQUESTS_PROXY_NON_DEBRID_ENABLED`	`true`	Set to `false` to bypass the proxy for general HTTP calls (catalog, indexers, content discovery). Debrid provider routing is unaffected.
`SCRAPLING_PROXY_URL`	`None`	Separate proxy for browser-based scraping
`TCP_KEEPALIVE_SECS`	`15`	TCP keepalive interval for all outbound HTTP clients (seconds). Keeps the proxy tunnel's NAT/conntrack mappings alive during idle periods.
`EGRESS_WATCHDOG_ENABLED`	`true`	Restart the pod when sustained egress loss is detected (see env reference).

Scheduler control¶

Variable	Default	Description
`DISABLE_ALL_SCHEDULER`	`false`	Disable all background scheduling (useful during development)
`TASKIQ_SINGLE_WORKER_MODE`	`true`	Route all task queues to one worker