Deploying a Python bot (aiogram, discord.py, FastAPI) to production

Prepare the project

Pin dependencies in requirements.txt.
Read tokens with os.environ, never hard-code them.
Log to stdout so the host shows logs.
Add a Dockerfile (our templates include one).

FROM python:3.12-slim
WORKDIR /app
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
COPY . .
USER nobody
CMD ["python", "-m", "bot.main"]

Deploy

Railway: connect the GitHub repository, add the variables, deploy. Tutorial: deploy a bot on Railway.

Vultr: create a venv on the server and a systemd unit. Tutorial: bot on a Vultr VPS with systemd.

Templates: aiogram · discord.py · FastAPI WhatsApp

Options we use and recommendAffiliate links: if you sign up, BotNetwork may earn a commission at no extra cost to you. See all recommendations

VPS / cloud

Vultr

A cloud provider with small VPS plans in many regions. You get full root access and predictable monthly pricing; updates, firewall and backups are your responsibility.

Best for: always-on bots where you want full control.

OfferNew users can get up to $300 in trial credit, per Vultr's promo terms.

Try Vultr

Affiliate link

Platform as a service

Railway

A deployment platform: connect a GitHub repository or Dockerfile and Railway builds and runs it, with logs, environment variables and managed databases. Usage-based pricing, so check how always-on services add up.

Best for: gateway bots and workers without managing a server.

OfferNew users get $20 in credits.

Try Railway

Affiliate link

Frequently asked questions

Which Python version?

Our templates require Python 3.10 or newer; use a currently supported release.

Do I need Docker?

No, but it makes the build reproducible on any host.

How do I keep it running after crashes?

Use the platform restart policy or systemd Restart=always.