standalone

@muleta-dev/standalone

Runnable muleta binary. Reads environment variables, binds a port, serves the JSON API.

Not published — this is the consumer of @muleta-dev/core and @muleta-dev/server that turns them into a running process.

Run the pre-built image (fastest)

Multi-arch image on GHCR — works on linux/amd64 + linux/arm64 (Apple Silicon). Point it at any reachable Redis:

docker run --rm -p 3737:3737 \
  -e MULETA_REDIS_URL=redis://host.docker.internal:6379 \
  ghcr.io/muleta-dev/muleta:edge

Then open http://localhost:3737.

Available tags:

Tag	Moves	Use for
:edge	every push to main	early adopters, always-latest
:latest	newest stable release	production
:X.Y.Z	pinned forever	reproducible deploys
:sha-<short>	pinned forever	debugging a specific commit

Run with Docker Compose

Starts Redis + muleta together; builds from source so you can tweak the Dockerfile:

cd apps/standalone
docker compose up --build

Then from another terminal:

curl http://localhost:3737/api/v1/queues

Run from source

Useful during development:

# 1. Start a Redis
docker run --rm -p 6379:6379 redis:7-alpine

# 2. Boot muleta (from the repo root)
MULETA_REDIS_URL=redis://localhost:6379 \
pnpm -F @muleta-dev/standalone dev

# 3. Hit the API
curl http://localhost:3737/api/v1/queues

Environment

Copy .env.example to .env and fill it in.

Variable	Required	Notes
MULETA_REDIS_URL	yes	redis://[:password@]host:port[/db]

The standalone always binds 3737. If you need a different port, map it at the proxy / container layer (docker run -p 8080:3737).

Queues are auto-discovered from Redis — any BullMQ queue under the default bull: prefix shows up within 15 s. No registration step required.

Shutdown

Sends SIGTERM or SIGINT (Ctrl-C). The process drains in-flight requests, closes the Redis connection, then exits.