How to Deploy a Flask App to Production
Flask's built-in server is for development only. How to deploy a Flask app properly: the app factory and config from environment variables, Gunicorn, a reverse proxy with HTTPS, running it as a systemd service, a Docker option, and the mistakes that cause 502s.
When you run flask run (or app.run()), Flask prints:
WARNING: This is a development server. Do not use it in a production deployment.
Use a production WSGI server instead.
Take it at its word. Here's how to deploy properly. (Flask vs Django vs FastAPI)
The production setup
Browser → Caddy/Nginx (HTTPS) → Gunicorn (several workers) → your Flask app
Gunicorn is a production WSGI server: it runs multiple copies of your app to handle requests in parallel and restarts workers that crash.
Step 1: configuration from the environment
Don't hardcode secrets or debug mode:
# app.py
import os
from flask import Flask
def create_app():
app = Flask(__name__)
app.config["SECRET_KEY"] = os.environ["SECRET_KEY"]
app.config["SQLALCHEMY_DATABASE_URI"] = os.environ["DATABASE_URL"]
# register blueprints, extensions...
return app
app = create_app()
- Never run with
debug=Truein production — the debugger lets anyone who sees an error page run Python code on your server. SECRET_KEYsigns session cookies; it must be long, random and secret. (Environment variables explained)
Step 2: dependencies
pip install gunicorn
pip freeze > requirements.txt
Step 3: try Gunicorn locally
gunicorn --bind 127.0.0.1:8000 --workers 3 app:app
app:app means "the app object in app.py". A common rule of thumb is 2–4 workers per CPU core.
Step 4: on the server
git clone https://github.com/you/myflask.git /srv/myflask
cd /srv/myflask
python3 -m venv .venv
.venv/bin/pip install -r requirements.txt
Create /srv/myflask/.env with your secrets, readable only by the app user (chmod 600). (Python virtual environments)
Step 5: run as a systemd service
# /etc/systemd/system/myflask.service
[Unit]
Description=My Flask app
After=network.target
[Service]
User=app
WorkingDirectory=/srv/myflask
EnvironmentFile=/srv/myflask/.env
ExecStart=/srv/myflask/.venv/bin/gunicorn --bind 127.0.0.1:8000 --workers 3 app:app
Restart=always
[Install]
WantedBy=multi-user.target
sudo systemctl daemon-reload
sudo systemctl enable --now myflask
journalctl -u myflask -f
Step 6: HTTPS
With Caddy:
example.com {
reverse_proxy 127.0.0.1:8000
}
Caddy gets the certificate automatically. (What is Caddy?) With Nginx, use proxy_pass http://127.0.0.1:8000; and Certbot. (Nginx reverse proxy config)
Tell Flask it's behind a proxy so it sees the real scheme and client IP:
from werkzeug.middleware.proxy_fix import ProxyFix
app.wsgi_app = ProxyFix(app.wsgi_app, x_for=1, x_proto=1)
Alternative: Docker
FROM python:3.13-slim
WORKDIR /app
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
COPY . .
USER nobody
CMD ["gunicorn", "--bind", "0.0.0.0:8000", "--workers", "3", "app:app"]
Note 0.0.0.0 inside a container. (Dockerfile explained)
Troubleshooting
- 502 Bad Gateway — Gunicorn isn't running or isn't on the port the proxy expects. Check
journalctl -u myflask. (502 Bad Gateway) - Worker timeout — a request took longer than Gunicorn's default 30 seconds. Make it faster or move the work to a background job. (504 Gateway Timeout)
- ModuleNotFoundError — the service uses a different Python than where you installed packages; point
ExecStartat the venv's gunicorn. (ModuleNotFoundError) - Static files missing — serve
static/from the proxy, or use WhiteNoise.
EasySpawn runs your Flask app behind HTTPS with Gunicorn, Postgres and daily backups set up — push your code and it's live. See how it works or join the waitlist.
Related: Flask vs Django vs FastAPI · How to Deploy a Django App · Deploy a FastAPI App to Production · How to Run a Python Script 24/7
Keep reading
How to Deploy a Streamlit App: Community Cloud or Your Own Server
Two good ways to put a Streamlit app online: Streamlit Community Cloud for free public apps straight from GitHub, or your own server with systemd and Nginx for private apps, more resources and no sleeping. Step-by-step for both, including secrets, requirements.txt and the WebSocket setting that trips people up.
What Is Vercel? What It Does, What It Costs You, and When to Use Something Else
Vercel is a hosting platform built around frontend frameworks, especially Next.js: push to GitHub and your site is live. What it actually does, what serverless functions mean for your app, where the limits are, and when a regular server is a better fit.