Documentation/Administration

Self-host

The same Docker image as the hosted version. Run it in your own AWS, GCP, or on-prem environment with your own PostgreSQL, S3, and SMTP.

System requirements

CPU: 2 vCPU for the API and 1 for the worker, minimum.
RAM: 4 GB for the API + 2 GB for the worker, minimum.
PostgreSQL: 16+ with the pgvector extension (optional, only needed for AI).
Redis: 7+ for queues and caching.
S3-compatible storage: AWS S3, GCS, MinIO, Wasabi.
SMTP: Any SMTP provider, or the Resend/SES API.

Running it with Docker Compose

# Log in to get access to the image
docker login ghcr.io -u {your-username}

# Download the compose file
curl -O https://raw.githubusercontent.com/pegalio/deploy/main/docker-compose.yml

# Create a .env with DATABASE_URL, REDIS_URL, S3_*, SMTP_*
cp .env.example .env
nano .env

# Bring it up
docker compose up -d

Environment variables

The full list lives in .env.example. At minimum:

DATABASE_URL — the PostgreSQL connection string.
REDIS_URL — the Redis connection string.
S3_* — endpoint, region, bucket, access keys.
SESSION_SECRET — at least 32 characters.
BASE_URL — your installation's public URL.
SMTP_* — for magic links and notifications.

Production

In production, make sure to change SESSION_SECRET to a long random value. The default is rejected in production mode and the server won't start.

HTTPS and a reverse proxy

Put a reverse proxy (Nginx, Caddy, Traefik) in front of the container. Caddy will issue a Let's Encrypt certificate automatically:

firstvalue.your-company.com {
  reverse_proxy localhost:3001
}
onboard.your-company.com {
  reverse_proxy localhost:3002
}

Backups

It's enough to back up PostgreSQL and the S3 bucket. Redis is just cache and queues — you don't need to back it up (queues recover; the cache just rebuilds).

Updates

Releases ship every sprint, patches weekly. Use semantic tags (:0.42, :latest). Database migrations run automatically when the container starts.