Astral.sh (the company behind uv) paid product: is it going to be a Heroku replacement?

Posted by DouweB82@reddit | Python | View on Reddit | 8 comments

As you might know Astral, the company behind uv and ruff, are a small company but have venture capital funding. And right now they are not making money at all (but of course they're building awesome tools!)

I listened to the Talk Python podcast episode 476 where the host & Charlie Marsh discuss the new capabilities uv has.

And it got me thinking, since uv is now able to install specific python versions, and since they stated they don't want to charge for features in uv ever, but are planning on charging for features adjacent to it that you might need to get your code up and running, could it be they'll be building a Heroku replacement? Certainly when you're wanting to build a Heroku-like thing it can be beneficial to allow users to specify their exact python version and have ways to install it.

I think it might be very cool and I'm sure there's a lot of opportunity in that space.

Does anybody know if Charlie Marsh or other Astral people ever discussed their plan to make money in more detail?

[-]

ExternalUserError@reddit

Just idle speculation on my part:

Free tools with telemetry that can be valuable to sell
Free tools that integrate with paid enterprise tooling
Free tools that eventually have a paid version
They eventually go Redis-style and enshittify

DanCardin@reddit

I think they’d make money stealing customers from aws/jfrog for a cheaper/better hosted private pypi index. Artifactory is insanely expensive and my company would switch off it in a second for something cheaper and with an easy migration option

zurtex@reddit

I would reccomend having general discussions with them on their Discordm there's an Astral > General category.

Though don't be surprised if they don't want to go into specifics about their business plan.

_Answer_42@reddit

Probably they will try to mirror anaconda.com

Anaconda were able to leverage their position from a free offering to a paid situation by doing two things:

Creating and hosting an independent ecosystem that specific sectors (data science) started to depend on
Publishing curated bundled environment with some additional features on top

I don’t see astral following this same business model for a number of reasons, some reasons include: Building up that ecosystem took many years and relied on creating a whole new packaging system that was not equivalent to Python’s existing system. A lot of bad will was created in the free to paid transition, and people are now far more wary of companies pulling the same trick.

Now, there are a few things Anaconda does that Astral could do and sell. For example a development creation environment that involves a GUI for non-developer users (data scientists, traders, etc.), when I maintained an enterprise copy of Anaconda I was surprised how many users relied on the Anaconda Navigator GUI even though it had terrible performance and was really clunky.

Matanya99@reddit

I just hope they stick around, they are accidentally turning python into my go-to for basically anything that isn't latency sensitive.

tocarbajal@reddit

The same business model as devbox from jetify, a great dev environment management tool with the exception that devbox has no free tier.

DouweB82@reddit (OP)

Ah, I see they have 'jetify cloud' , yeah I can see how that could be similar