Square CloudSquare Cloud

Python ready for production

FastAPI, Django, Flask or an AI worker on Python 3.14. pip install reads requirements.txt or pyproject.toml, and ffmpeg and headless Chromium are already in the image.

From ZIP to running process

The same path on every deploy, from the dashboard, the CLI or GitHub.

1

Detects the language

RUNTIME, or the extension of MAIN (.py), picks Python.

2

Installs the dependencies

requirements.txt is read and the dependencies are installed on the server, before the app starts.

3

Isolates the container

The CPU and RAM of MEMORY are guaranteed by the kernel, and the process runs without root.

4

Starts the process

Without START, the default command runs. With START, yours runs exactly as written.

The squarecloud.app of a Python app

A KEY=VALUE file at the root of the project, or generated by the dashboard at deploy. This one publishes a site or API; without SUBDOMAIN, the app runs as a background process.

squarecloud.app
  1. 1MAIN=main.py
  2. 2MEMORY=512
  3. 3VERSION=recommended
  4. 4DISPLAY_NAME=my-api
  5. 5SUBDOMAIN=my-api
  6. 6START=python -m uvicorn main:app --host 0.0.0.0 --port 80
  7. 7AUTORESTART=true
Versions
3.14
Default command
python main.py
Dependencies
requirements.txtpyproject.toml
Accepted files
.py
RUNTIME= values
python
Also in the image
Node.js 26ffmpegChromium

What comes ready

The environment already knows your language: you upload the code and it runs.

requirements.txt or pyproject.toml

Use the format your project already has: the classic requirements.txt or the standard pyproject.toml.

Faster installs

Pre-built wheels take priority, which shortens installs for libraries with native extensions.

AI and ML libraries

torch, transformers, sentence-transformers, and the OpenAI SDK work with no extra setup, within your plan's RAM.

Native asyncio

asyncio and async libraries like aiohttp and httpx work with no additional configuration.

What you can host

Projects that run on this runtime.

APIs with FastAPI or Flask

Async backends with uvicorn, Pydantic, and SQLAlchemy, from prototype to production.

Scrapers and crawlers

Data collection with requests, httpx, BeautifulSoup, Scrapy, or Playwright in headless mode.

AI model inference

Models with transformers and sentence-transformers, or apps built on the OpenAI SDK and LangChain.

Workers, queues, and ETL

Celery or RQ tasks and data pipelines with pandas and polars running in the background.

Full Django applications

Django projects with the ORM, the built-in admin, and server-rendered templates.

Discord and Telegram bots

Bots online 24/7 with discord.py, aiogram, python-telegram-bot, or Telethon, with auto-restart.

Combine with other services

Everything in the same dashboard, on the same infrastructure and the same invoice.

A plan for every stage of your project

From your first deploy to 1 TB of RAM, without switching platforms. Each phase adds memory and features to the one before, and upgrades apply instantly.

Compare all plans
1 GB → 2 GB

Hobby

Take the idea off localhost and put it online.

  • Up to 2 vCPU
  • Real-time metrics
  • Blob Storage
  • One-click rollback
  • Square Shield anti-DDoS
  • Cubic AI agent

FromR$12.49/mo

4 GB → 8 GB

Standard

Recommended

Production with databases, your own domain and a team.

  • Up to 4 vCPU
  • Databases
  • Custom domain
  • Load balancer
  • AI Gateway
  • Workspaces

FromR$44.99/mo

12 GB → 24 GB

Pro

More memory, daily snapshots and visibility into what runs.

  • Up to 12 vCPU
  • Daily application snapshots
  • Performance analytics
  • Cubic Pro and routines
  • Priority support

FromR$124.99/mo

32 GB → 1 TB

Enterprise

Large operations, with a private Discord channel.

  • Up to 16 vCPU
  • Private Discord channel

FromR$299.90/mo

HobbyStandardProEnterprise
Compute
RAM1 GB → 2 GB4 GB → 8 GB12 GB → 24 GB32 GB → 1 TB
vCPU1 → 246 → 1216
Platform
Real-time metricsIncludedIncludedIncludedIncluded
Blob StorageIncludedIncludedIncludedIncluded
One-click rollbackIncludedIncludedIncludedIncluded
Square Shield anti-DDoSIncludedIncludedIncludedIncluded
Cubic AI agentIncludedIncludedIncludedIncluded
DatabasesNot includedIncludedIncludedIncluded
Custom domainNot includedIncludedIncludedIncluded
Load balancerapps per domainNot included2510
AI GatewayNot includedIncludedIncludedIncluded
Daily application snapshotsNot includedNot includedIncludedIncluded
Performance analyticsNot includedNot includedIncludedIncluded
Cubic Pro and routinesNot includedNot includedIncludedIncluded
Team and support
Workspacesmembers per workspaceNot included3 → 710 → 2020 → 50
Priority supportNot includedNot includedIncludedIncluded
Private Discord channelNot includedNot includedNot includedIncluded

4.9 out of 5

482 reviews on Google and Trustpilot

Hosted in a partner datacenter with certifications

  • Tier 3
  • ISO 27001
  • SOC 2 Type II
  • Always-On DDoS Protection
Explore our infrastructure

Frequently asked questions about Python Hosting

Versions, dependencies, ML libs, and Python bots.

Which Python versions are available?

Latest and Recommended: Python 3.14. Both tracks ship the same version. Set via VERSION=.

Can I use pyproject.toml instead of requirements.txt?

Yes. The platform accepts both, so pick whichever format you prefer (PEP 621 or classic). Deps are installed before start.

Do AI libraries like torch and transformers work?

Yes. The installer prefers pre-built wheels, saving minutes on boot. torch, transformers, sentence-transformers, and similar run out of the box.

Which bot framework do you recommend?

For Discord: discord.py or nextcord. For Telegram: aiogram, python-telegram-bot, or Telethon. All are supported.

Ship your project today

Join the 500,000+ developers who have used Square Cloud to host their projects.

  • Deploy in seconds via GitHub, CLI or .zip
  • Isolated, always-on containers
  • Automatic TLS on every domain
  • Anti-DDoS on layers 3, 4 and 7
  • Snapshots with 1-click restore
  • Cubic, the AI agent, on every plan
  • Fixed price in reais, with NFS-e
  • No per-request charges