Configuration

Environment variables

The compose file handles all internal service connectivity automatically. The only variables you configure in your .env are the ones below. REDIS_URL and MEILISEARCH_HOST default to the correct Docker-internal hostnames and never need to be set in production.

Required

Variable	Description
`BETTER_AUTH_SECRET`	Secret key used to sign sessions and tokens. Generate with `openssl rand -base64 32`.
`POSTGRES_PASSWORD`	Password for the internal PostgreSQL database. Choose a strong password.
`MEILISEARCH_MASTER_KEY`	Meilisearch master key. Generate with `openssl rand -hex 32`.
`TMDB_READ_ACCESS_TOKEN`	TMDB v4 Read Access Token for movie and TV metadata enrichment. Get one free at themoviedb.org.

Optional

Variable	Description
`PASSKEY_RP_ID`	Your deployment's public hostname, e.g. `minato.example.com`. Required to enable passkey (WebAuthn) support. Passkeys are disabled when this is not set. See Passkeys.
`OPENROUTER_API_KEY`	OpenRouter API key. Only needed if you enable AI-powered metadata repair in Settings. See AI Metadata repair.

Generating secrets

# BETTER_AUTH_SECRET
openssl rand -base64 32

# MEILISEARCH_MASTER_KEY
openssl rand -hex 32

Don't use passwords you've used elsewhere. These values are security-sensitive — if either leaks, rotate them and restart the stack.

Rotating a secret

To rotate BETTER_AUTH_SECRET, update it in your .env file and restart the stack. Existing user sessions will be invalidated and users will need to log in again.

To rotate MEILISEARCH_MASTER_KEY, update it in your .env file and restart the stack.

Volumes

All state lives in Docker named volumes. Back these up to avoid data loss.

Volume	What it holds
`minato_config`	Poster and backdrop images (`/config/media`) and community scrapers (`/config/scrapers`)
`postgres_data`	The PostgreSQL data directory — your entire torrent database
`redis_data`	Redis persistence for the BullMQ job queue
`meilisearch_data`	Meilisearch index data

The Meilisearch index can be rebuilt from PostgreSQL if lost — it's not the source of truth. The other three cannot be reconstructed, so treat them as critical data.

Backing up

# Postgres — take a logical dump
docker compose exec postgres pg_dump -U minato minato > minato_$(date +%Y%m%d).sql

# Full volume backup (works for any volume)
docker run --rm \
  -v minato_config:/data \
  -v $(pwd):/backup \
  alpine tar czf /backup/minato_config_$(date +%Y%m%d).tar.gz -C /data .

Restoring Postgres

cat minato_backup.sql | docker compose exec -T postgres psql -U minato minato

Ports

Only one port needs to be exposed to the host.

Port	Service
`7271`	nginx — serves the web frontend and proxies `/api/` requests to the Hono backend

All other services (PostgreSQL, Redis, Meilisearch) communicate on the internal Docker network and are not exposed. Don't expose them unless you have a specific reason to.

Community scrapers

Community scrapers are loaded from /config/scrapers inside the container, which is part of the minato_config volume. To use a host directory instead (so you can drop in scraper packages without touching Docker volumes), bind-mount it:

services:
  minato:
    volumes:
      - minato_config:/config
      - ./scrapers:/config/scrapers   # add this line

Drop scraper directories into ./scrapers/ on your host. Minato picks them up automatically without a restart.

Runtime settings

Several instance-wide settings are configured through the admin dashboard (Settings) rather than environment variables. They are stored in the database and applied at runtime without a restart.

Search ranking

Profile	Behaviour
Quality	Prefers well-seeded torrents with high-resolution quality indicators.
Health	Prioritises torrents with the most seeders and fewest leechers.
Freshness	Surfaces recently ingested content first.

Set the profile under Settings → Search → Ranking profile.

Worker concurrency

Setting	Range	Default	Description
Ingest concurrency	1–50	5	Parallel jobs parsing release titles and indexing in Meilisearch.
Enrichment concurrency	1–20	5	Parallel jobs fetching metadata from TMDB and AniList.

Higher values increase throughput but also load on external APIs (TMDB rate limits). Adjust these under Settings → Workers.