Deployment

Radhflow ships as a single Docker image. The same image runs on your laptop, on EU cloud infrastructure, or on a customer’s on-premise server. No code forks. No edition-specific features. Configuration determines the mode.

Three modes

Local

Your machine. Your data. Your Git repo.

docker run -p 8080:80 \
  -v ~/.radhflow/projects:/rf/project \
  -v ~/.radhflow/config:/rf/config \
  ghcr.io/radh-io/radhflow:latest

Data stays on disk. Pipelines live in a local Git repository. No cloud account required. No telemetry. Credentials stored in local SQLite.

Best for: development, personal automation, sensitive data processing.

SaaS (hosted)

EU container infrastructure. Managed scheduling. Credential vault.

Radhflow SaaS runs on Hetzner and Scaleway — EU-sovereign providers. Data stays in EU jurisdiction. Pipeline artifacts stored in customer-scoped S3 buckets on EU endpoints.

Component	Infrastructure
Compute	Hetzner Cloud (Nuremberg / Helsinki)
Object storage	Scaleway S3 (Paris / Amsterdam)
Container registry	GitHub Container Registry
DNS + CDN	Cloudflare (EU-only regions)

No US hyperscaler in the data path. No customer code stored beyond execution.

Best for: teams that need scheduling, shared credentials, and managed infrastructure without US cloud dependency.

Enterprise

Customer VPS. Customer data. Customer Git.

The same Docker image runs on the customer’s infrastructure. Radhflow provides the image; the customer provides compute, storage, and Git hosting. No data leaves the customer’s network.

# Customer's infrastructure
docker compose up -d

Configuration points to the customer’s Git server, storage backend, and credential store. Radhflow has no access to the customer’s instance.

Best for: regulated industries, air-gapped environments, organizations with existing infrastructure.

Environment variables

Variable	Default	Description
RF_PORT	80	HTTP port inside the container
RF_PROJECT_PATH	/rf/project	Pipeline workspace root
RF_CONFIG_PATH	/rf/config	Config and database directory
RF_MODE	local	Deployment mode: local, saas, enterprise
RF_LOG_LEVEL	info	Log verbosity: debug, info, warn, error
RF_S3_ENDPOINT	—	S3-compatible endpoint for artifact storage
RF_S3_BUCKET	—	Bucket name for pipeline artifacts
RF_S3_REGION	eu-central-1	S3 region
RF_GIT_REMOTE	—	Remote Git URL for pipeline sync
RF_SCHEDULER	false	Enable cron-based pipeline scheduling
RF_CREDENTIALS_KEY	—	Encryption key for credential vault

Docker Compose

Full local setup with persistent volumes:

version: "3.9"
services:
  radhflow:
    image: ghcr.io/radh-io/radhflow:latest
    ports:
      - "8080:80"
    volumes:
      - ~/.radhflow/projects/default:/rf/project
      - ~/.radhflow/config:/rf/config
    environment:
      RF_MODE: local
      RF_LOG_LEVEL: info
    restart: unless-stopped

docker compose up -d

The web UI is available at http://localhost:8080. The API is at http://localhost:8080/api.

Fly.io deployment

For hosted or team deployments, Fly.io provides EU regions with persistent volumes:

fly launch --image ghcr.io/radh-io/radhflow:latest \
  --region ams \
  --vm-size shared-cpu-1x

fly secrets set RF_MODE=saas \
  RF_S3_ENDPOINT=https://s3.fr-par.scw.cloud \
  RF_S3_BUCKET=my-team-artifacts \
  RF_CREDENTIALS_KEY=$(openssl rand -hex 32)

fly volumes create rf_data --region ams --size 10
fly deploy

Port mapping

The container listens on port 80 internally. Map it to any host port:

# Development
docker run -p 8080:80 ghcr.io/radh-io/radhflow:latest

# Production (behind reverse proxy)
docker run -p 127.0.0.1:3000:80 ghcr.io/radh-io/radhflow:latest

Data persistence

Path	Contents	Backup strategy
/rf/project	Pipeline YAML, node specs, generated code	Git push
/rf/config/rf.db	Credentials, events, preferences	SQLite backup or volume snapshot
S3 bucket	Pipeline artifacts, NDJSON outputs	S3 versioning

Git is the primary backup mechanism for pipeline definitions. The SQLite database stores runtime state and credentials. For hosted deployments, S3 stores pipeline artifacts with versioning enabled.