Blog

FastAPI Labs

Blog

FastAPI Labs

FastAPI Cloud - By The Same Team Behind FastAPI

May 5, 2025

TL;DR: Imagine deploying your FastAPI apps with just a single command. I started a company to build FastAPI Cloud, so you can do exactly that.

Check it out and join the waiting list! 🚀

Deploy to FastAPI Cloud

If you have a FastAPI app, you can now run:

And that's it, we handle the rest. 😎

You get automatic deployments, HTTPS, autoscaling (including down to zero to save on costs), and more features to come, all optimized for FastAPI and Python.

You code. We cloud.

Go ahead and join the waiting list, we'll gradually let people in during the beta.

Why FastAPI Cloud

So, why would you want to use FastAPI Cloud?

For the same reasons that you want to use FastAPI instead of something else. 😅

Why Use FastAPI

You could build web apps and APIs in other ways, but you probably like what FastAPI gives you:

Great developer experience
Velocity building your app
Certainty that you're doing things right
And more... [think of what you like from FastAPI and add it here]

For example, in FastAPI, by using Pydantic validation, you know your code and your data are correct while staying simple. And because the core of Pydantic is written in Rust, it gives you very good performance, while your app logic can stay in a much simpler language that lets you build faster: Python.

Your app won't crash or, even worse, silently do something wrong, just because a client sent incorrect data, it’s all validated. You're already safe from a good range of accidents and even some potential attacks.

You don't need to write an API definition by hand with OpenAPI (can you imagine having to do that?). It's already there, autogenerated from your code, by default. You get automatic API docs.

And you don't even need to think about any of that, you're probably even used to all these automated features by this point.

Why Use FastAPI Cloud

Now let's think about the cloud.

You can (and will always be able to) deploy FastAPI in any cloud you want. But deploying anything to the cloud is normally a lot of work. For example, check the FastAPI docs explaining the concepts you should understand to think about deployments in general, they're quite big, even though I did try to keep them as minimal as possible.

Note: I still encourage you to read the docs, but I know not many will. 😅

The Problem

Once you want to have something more robust, serve real users, handle scale, failures, etc., your main option is to learn Kubernetes, and understand the “cloud native” tools you would need sitting on top, plus some cloud providers and their many available products.

Or maybe you would need to tweak one of the cloud products that try to abstract that complexity away in different ways, but they are often not optimized for Python. You would have to figure out how to do it all by trial and error, which involves lots of frustration and uncertainty in the final result.

You spend months with all that, just to discover that the whole area is always evolving, and keeping pace is hard.

After all this effort, you have a way to deploy your app that you think is probably relatively good, maybe. There's a lot of uncertainty, and a lot of things can be done incorrectly, even if it seems to work. And that's an annoying feeling.

The Solution

So, you went and learned FastAPI and built a product that is pretty much ready, all in a couple of weeks... and you only need this tiny final step to let users try it: deploy it to the cloud. Your product is ready, but this tiny step would require months of study and work, filled with frustration, to yield something that seems to work, but still leaves you with uncertainty.

That's what we are solving, we make that tiny step of deploying to the cloud, actually tiny for you.

We move mountains behind the scenes so you can just:

And you can have the same level of certainty in your cloud that you get from your framework.

The Right People

Now, anyone can promise you all that.

But as you should do with politicians, don't look at promises, look at statistics, results, and facts you can check.

GitHub stars are a bad proxy for popularity, and still the best one we have so far. 😅 By that count, I made for you the most starred:

Web framework (across languages): FastAPI - 84k+ ⭐️
SQL library (ORM) in Python: SQLModel - 15k+ ⭐️
Command Line app library in Python: Typer - 16k+ ⭐️

I hope you know that my team and I know Python. 😉 We have always worked hard on giving you the best Python developer experience. Now we’ll also do it for your cloud.

If you've hung around my open source tools before, you might have bumped into some of us already, like:

You know we can ensure that FastAPI Cloud works well, that it's fast, secure, has the right things, and has all you would expect by now.

And as you're working in Python, there's a big chance you're building with AI, or that you will at some point. Rest assured, we'll have your back then, too. I have old, strong roots in AI and ML, that's how I got into Python 10+ years ago, I just got stuck in solving APIs. 😅

This is the cloud for FastAPI, and in extension, for Python, including AI, by the team behind FastAPI.

Don't Worry, Cloud Happy

We're all developers here, we learn to be skeptical, and at best, cautiously optimistic.

So let's see what could be there to worry about, and why there's nothing to worry about. This way, you can stay cautiously optimistic. 😉

Open Source

What does this mean for my open source tools?

I’m starting a company, so you just have some extra certainty that the projects will stay maintained and working for a very long time.

Now it's not just my willpower trying to accommodate everything to do it on my own. There's a company that pays engineers to work on open source projects.

The more people use FastAPI and Python successfully, the more potential clients for the product. So, good open source is good business for the company. It’s all nicely aligned. 😎

Vendor Lock-in

TL;DR: I hate vendor lock-in, it won’t happen.

One of the first things I'm cautious about when trying a cloud product is the vendor lock-in.

When a product is set in a way that, once you start using it, if you stop, you would lose your data, or the effort you invested in writing your code, etc., you're "locked" in that vendor. That's a sensible concern to have. I don't like that and I run away from it when I can.

FastAPI is open source, based on open standards, you can already run it on your own on any cloud you want. I probably have more docs about it than any other framework. Those docs will stay there. And being able to deploy FastAPI however you want will stay like that. Additionally, I spent years of my life earning your trust, even taking pay cuts and rejecting good offers just to work on open source; I would not risk that trust and reputation now.

I hear companies talk about moats, but to me, that sounds as if we were in the Middle Ages, with castles, and imaginary dragons... and maybe horses made of coconuts. I don't think building extra walls or restrictions to cross has ever helped much (and dragons can fly over moats anyway).

I think one of the best "moats" is an open door.

Have you ever seen a kid being hugged tightly by some annoying relative who wouldn't let go? That's probably the situation where you saw that kid want to leave the most.

Now, put the kid in a room with an open door so the kid can leave whenever they want, and provide the best toys available, the ones the kid enjoys the most, in that room. That kid would happily stay there for even longer than you wanted.

That's my intention, you can use any other cloud you want, and you will be able to deploy FastAPI however you want. If you already solved these problems a long time ago and don't want to touch that after so much effort, that's perfectly fine too, you’ll still reap the benefits of the improvements in FastAPI and the open source ecosystem.

We'll make sure we make the developer experience using FastAPI Cloud the best you can get in Python, so you’ll hopefully stay not because you can't leave, but because you don't want to leave.

How to Cloud

Okay, enough talk about “why” this or that. How do you use FastAPI Cloud?

First, make sure you join the waiting list. Go ahead, I’ll wait. ☕️

Let's say we already let you in, and let’s say you have a FastAPI app in a file main.py:

from fastapi import FastAPI


app = FastAPI()


@app.get("/")
def main():
    return {"message": "Hello World"}

You install FastAPI, for example:

$ uv add "fastapi[standard]

Activate the virtual environment:

Start the FastAPI server (which uses Uvicorn) in dev mode:

You can open your browser at that address and confirm it is all working correctly.

Note: You can also run the production server with fastapi run instead of fastapi dev.

If that's working locally, you can now deploy your app to FastAPI Cloud (after we let you in from the waiting list).

First, login:

Next, deploy:

That's it! 🤯 You get a URL with your app deployed and ready. 😎

Note: extra points if you figure out the Colombian chicken joke.

We handle building and deploying your app, HTTPS, autoscaling/replication based on requests, even scale down to zero, to save you costs, all optimized for FastAPI and Python.

Later, you can also configure environment variables for your app, integrate with databases, custom domains, etc.

Backers

We are very fortunate to have the best support anyone could ask for.

We raised a seed round of $4.2M, led by Sequoia, plus several angel investors.

One could say a good part of the Python and AI communities are already supporting us since the inception of the idea, some of our angels and advisors include:

Yury Selivanov
- CPython core developer, asyncio, uvloop, asyncpg, Gel
Samuel Colvin
- Creator of Pydantic
Paige Bailey
- Google DeepMind
Jessie Frazelle
- Docker, Google, GitHub, Oxide, Zoo
Carol Willing
- CPython and Jupyter core developer
Tom Christie
- Creator of Django REST-Framework, Starlette, Uvicorn, HTTPX
Travis Oliphant
- Creator of NumPy
Charlie Marsh
- Creator of Ruff, uv, founder of Astral
Armin Ronacher
- Creator of Flask, Click, Jinja2
David Cramer
- Founder of Sentry
Peter Wang
- Co-founder of Anaconda
Clem Delangue
- Co-founder of HuggingFace
Marcelo Trylesinski
- Pydantic, maintainer of Starlette, Uvicorn
Martin Donath
- Creator of Material for MkDocs
Robert Nishihara
- Co-founder of Anyscale, creator of Ray
Birk Jernström
- Founder of Polar.sh
Sahir Azam
- CPO of MongoDB
Alex Pall
- Mantis VC, The Chainsmokers

Build Harder, Better, Faster, Stronger

There are so many app ideas I've had over the years that I never built because deploying them was so complex. And I actually already knew how to do it, but still, doing it was so complex.

With FastAPI Cloud, I can now finally build all those ideas, from the silliest quick apps to the more sophisticated ones.

And you can now do the same, you can create new quick random things, create giant apps and products, or migrate existing apps just to get a better value from your cloud and more peace of mind.

I'm excited to use FastAPI Cloud myself to build stuff.

And I would love for you to feel that new superpower too.

Go build stuff, build anything, build everything! 🚀