Deploy

Ship reflex-django in one of two ways. Option 1 uses Reflex production commands (reflex deploy or reflex run --env prod). Option 2 runs Django and Reflex as separate services with explicit plugin config.

Option 1: Reflex deploy (integrated)

One integrated stack: Django admin, API, SPA, and Reflex events. Set profile: "integrated" in production rxconfig.py:

# Integrated production rxconfig.py (Option 1)
import reflex as rx
from reflex_django.plugins import ReflexDjangoPlugin

config = rx.Config(
    app_name="shop",
    redis_url="redis://localhost:6379/0",
    plugins=[
        ReflexDjangoPlugin(
            config={
                "settings_module": "config.production",
                "profile": "integrated",
            }
        ),
    ],
)

Run migrations before every deploy:

export DJANGO_SETTINGS_MODULE=config.production
reflex django migrate --noinput

A. Reflex Cloud

Deploy to Reflex Hosting:

reflex deploy

Use a project config file if you have one (reflex deploy --config cloud.yml). Your Django settings and database must be reachable from the hosted environment.

B. Self-hosted (reflex run --env prod)

Production build and run on your own server:

reflex run --env prod

Browse the app on port 3000 (frontend) and 8000 (backend) by default. For a single backend port behind your own reverse proxy:

reflex run --env prod --backend-only --backend-port 8000

Set redis_url in rx.Config when you run multiple Reflex workers. Set RX_AUTO_EXPORT_ON_START=0 in Django settings so the app does not rebuild on every boot.

Before using the bundled fallback settings in production, read Security. Production runs warn when DEBUG, ALLOWED_HOSTS, SECRET_KEY, or cookie settings are unsafe.

When to pick Option 1

• Simplest path: one deploy command or one reflex run
• Single VPS, container, or Reflex Cloud
• Typical reflex-django apps

---

Option 2: Separate Django and Reflex

Two production services plus an edge reverse proxy:

Path	Service
/admin/, /api/, SPA shell (HTML)	Django
/_event, /_upload	Reflex backend

Django service config

Keep normal Django production settings (INSTALLED_APPS, AsyncStreamingMiddleware last). Build the SPA bundle once so Django can serve the shell:

export DJANGO_SETTINGS_MODULE=config.production
reflex django migrate --noinput
reflex export
reflex django collectstatic --noinput
uvicorn config.asgi:application --host 0.0.0.0 --port 8000 --workers 4

Django does not run reflex run in this layout. It serves admin, API, and the exported SPA via reflex-django mount.

Reflex service config

Turn embed off (Django is not inside the Reflex process). Keep mount and bridge on. Disable proxy (your edge proxy replaces dev Vite proxy):

# Split production rxconfig.py (Reflex worker - Option 2)
import os
import reflex as rx
from reflex_django.plugins import ReflexDjangoPlugin

config = rx.Config(
    app_name="shop",
    redis_url=os.environ["REDIS_URL"],
    plugins=[
        ReflexDjangoPlugin(
            config={
                "settings_module": "config.production",
                "embed": {"enabled": False},
                "mount": {"enabled": True},
                "proxy": {"enabled": False},
                "bridge": {"enabled": True, "mode": "full"},
            }
        ),
    ],
)

Run the Reflex backend:

export DJANGO_SETTINGS_MODULE=config.production
export REDIS_URL=redis://redis:6379/0
reflex run --env prod --backend-only --backend-port 8001

Keep mount.enabled: true on the Reflex service because the Reflex worker still needs the app's route/deep-link metadata and SPA/event wiring even though Django serves the public HTML shell in this layout. The edge proxy decides which service receives each request.

Reverse proxy

Route /_event (long WebSocket timeout) and /_upload to the Reflex upstream. Send everything else to Django.

Example: docs/examples/deploy/nginx.conf

Full layout: docs/examples/docker-compose.scaling.yml

Split checklist

1. embed.enabled: false in the Reflex service rxconfig.py
2. proxy.enabled: false in production
3. bridge.enabled: true so handlers keep self.request.user
4. Same DJANGO_SETTINGS_MODULE on both services
5. Shared REDIS_URL and cache-backed sessions (Scaling)
6. HTTPS cookie settings and stable SECRET_KEY (Security)

When to pick Option 2

• Scale Django and Reflex workers independently
• Django on an existing platform (Kubernetes, PaaS) plus a Reflex worker pool
• Different deploy cadence for API vs event backend

---

Docker sketches

Option 1 (integrated):

RUN reflex django migrate --noinput
CMD ["reflex", "run", "--env", "prod", "--backend-only", "--backend-port", "8000"]

Option 2: see docker-compose.scaling.yml (django + reflex + nginx + redis).

Commands

Command	Use
reflex run	Dev (integrated profile)
reflex run --env prod	Self-hosted production
reflex deploy	Reflex Cloud
reflex export	Build SPA static files (split Django service)
reflex django migrate	Database migrations

See Scaling for Redis and multi-worker tuning.

Next: Security and Troubleshooting