Thank you! Cirrus CLI is still around and can run your tasks locally in either Podman or Docker. Can also be used in any other CI.

I mostly clicked the link because I was curious if Cirrus Labs operates Cirrus CI and if so how that would be impacted.

Looks like I’ll need to move the FreeBSD CI jobs for open source projects I maintain to another solution. Anyone have suggestions for alternatives?

It's hard not to imaging that OpenAI is attempting to build an developer tools eco-system. It makes sense as it's one of the few fields in AI that are currently able to generate sales.

This is kind-of neat too, at least in the near term:

> In the coming weeks, we will relicense all of our source-available tools, including Tart, Vetu and Orchard under a more permissive license. We have also stopped charging licensing fees for them.

Just want to note that we will continue maintaining and improving our virtualization solutions actually with even greater attention. SaaS options like Cirrus CI and Cirrus Runners will eventually wind down so we can focus on incorporating pieces internally.

It could also be a suite of product acquisition, the CI could be a product OpenAI is interested in having, but not sell.

Is Sam or family an investor in them anyways?

This is a VC site. Morality or changing the world for the better is window dressing on earning as much money as possible and damn anyone who gets in the way. Morality is only good in as much as it’s good for business to be seen as moral. Money people have been majority running things for over 2 decades at least but the mythos lives on.

Cirrus CI was on a downhill in terms of users and revenue for years now. Most of the customers moved to GHA already.

Plus migration is super easy with Cirrus CLI -- tool to run our CI task definitions locally or in any CI. See https://github.com/cirruslabs/cirrus-cli

Realistically this is what you agree to when you want to use someone else's computer. They can just as easily ran out of money.

I don’t think they care about their customers at all, from the statement they consider their business a “byproduct”.

but they're not

> We are no longer accepting new customers for Cirrus Runners but will continue supporting the service for existing customers through their existing contract periods.

Same thoughts. Guess the migration team responsible will have to kick up the gear.

Ability to virtualize on Apple devices and linux with GPUs https://github.com/scipy/scipy/issues/24990

To create an ephemeral (docker-like) MacOS VM on a Mac with full performance and access (e.g. GPU) you have to use a virtualization API provided by Apple.

For most CI use, you can choose between:

  Anka, a "contact-us for pricing" closed-source projet, where you have to pay expensive license (easy 3000 USD/yr per machine)

or

  tart, which is a lightweight wrapper around the official Apple API.

But you have to know that on MacOS, there is an artificial limit of 2 VMs per Mac... but well:

https://github.com/cirruslabs/orchard/commit/3cfa2445500f45f...

With https://khronokernel.com/macos/2023/08/08/AS-VM.html

Some people might find it very attractive:

  Instead 25 Mac Mini you might need only 5.
  + No licensing to pay to Anka.

Even without bypassing the limit it is great actually

Better UX than their competitors and support for many different images.

That is something of a broken promise in the SaaS world; it was supposed to be convenience but instead it became a hundred broken integration points as you struggle to keep up with deprecated APIs and acqui-hires killing off shims you wouldn't have needed if everything had been on-prem to begin with.

Been there, done that, and I still find value in 3rd party services despite the occasional need to migrate.

Self hosting is the way to go if you need to keep monthly services spend as low as possible but you have extra time to spend, such as with a hobby project.

Whenever I’ve worked on real startup projects, self-hosting became a constant source of little tasks for the engineering team to mix into our weekly workload. There were always little tasks to upgrade this service, investigate why that one server was slow, or to migrate something to a bigger server because we were bottlenecked on some resource. Then we had to manage backups and do our recovery drills, along with changing the backup strategy every 6 months because someone had a better idea.

When we started to add up all of the time spent managing everything it starts to look like spending dollars (of engineer time) to save pennies on SaaS bills.

Probably not a popular thing to say on HN, but I now try to stay away from teams that go to extremes to self-host everything because I just want to get my work done, not also be constantly involved in running the underlying services. I do it for my own hobby projects at home but I don’t want to be doing it at work where we have money to spend to lighten the load. If the cost is the occasional migration to a different 3rd party service that’s not a big workload relative to everything involved in self-hosting.

This is the way.

More like employees number 32,463 and 32,464 - they’re two people from what I seen on GitHub over the years. (Incredibly strong two people)

> at this point, it appears that if one wants one's career to progress, you need to start some kind of tiny startup like Astral or Bun and hope to be notable enough you can get acquired by someone like OpenAI or Anthropic

This has been popular for 25+ years. Likely before, but that's when I first started noticing a significant number of companies that were clearly in business solely TO BE BOUGHT.

It also kills competition and disincentivizes providing long term value. I’m not sure how making the job market more like gambling does anything positive for the majority of people

bloating up for the IPO

> We have AI companies constantly fear-mongering that their next model is somehow too dangerous to release

I'm guessing you're referring to this recent report of the security vulnerabilities Mythos found and submitted patches for? That just seems like they don't want the negative press and/or liability if their new model ends up being used to create 0-days that cause widespread damage.

https://red.anthropic.com/2026/mythos-preview/

yeah super confused, it looks like some tools to manage vms in ci? what is unique about that vs lxc or docker or apple's native container cli?

FreeBSD CI testing is the part I’ll miss the most… time to find an alternative for open source projects.

“We will replace white collar workers”

Proceeds to buy white collar workers.

Same; the reason everyone ran out to buy Mac Minis last month is it gave their Claw access to iMessage, their browser cookies, and a residential IP. Cirrus provides a way to provision and orchestrate MacOS VMs, which is exactly what I did for running Openclaw (for a minute …).

They want to kill the last good ones

What software do you use to run your CI?

Making a statement like this is generally part of the terms of the acquisition.

It’s been true for a while, but AI seems to have exaggerated it. To me it reinforces the idea that LLMs created a “K” shaped talent market. Those whose are good become even more valuable.

Thank you! Full journey is too long for a comment here.

It was hard and I was lucky with my previous pre-IPO gigs at Airbnb and Twitter so I had some bootstrap fund. In retrospective a dev tools startup in 2017 with no network and no VC support was a crazy idea but I was young and didn’t think thought too much.

Then it was long 8 years of raw work and constant questioning this choice. Then finally a third component: luck. In 2024-2025 it kind of grew organically due to market changes and back in October 2025 I finally stopped questioning the future of Cirrus Labs.

My only advice if I may, try to get your first dollar from your startup while you are employed.

Cuirrus does not do AI, they do CI.

That part is amazing. Back when I first heard of tart I thought it was amazing, with the one downside being the license.

Hopefully development on it continues, or a community maintained version keeps it going.

This is making my current work 100 times easier. Very welcome timing.

This is a huge deal! I secretly hoped for this. :)

This is awesome because I flippin love tart.

Yeah. Much like Astral - acquiring both the product (because they need to use it internally, but don't care about trying to resell / market), and they also want the talent to keep maintaining it / add features they want.

For now.

We were 100% bootstrapped with no outside capital or support/advisory.

Killing bad people is changing the world for the better

If your scope includes making the Codex web app environments have additional functionality I look forward to it. More enterprise features and yaml backed pipelines.

What are your plans for tart licensing going forward?

https://github.com/cirruslabs/tart/blob/main/LICENSE

https://tart.run/licensing/

Part of my $DAYJOB is working on timers.

@sama if you need someone to buy to implement timers for ChatGPT I’m your guy - my price is 2 billion dollars.

not op either, github actions, self hosted runner on bare metal with lxc containers for each runner.

Cost savings are insane and the speed of latest amd epycs are miles ahead of the default ci instances on github and other places.

Not OP, but to me the answer is:

  - gitea
  - woodpecker CI
  - my own docker registry
  - portainer running on my docker swarm

I then define the docker stack in the git repository, and CI builds the images and pushes to build the new image to the docker repository. The portainer API allows to deploy a stack, with the image tag as a parameter.

That is good advice!

I am too late for that as I am full time but also lucky to have had a previous exit.

I am planning to go the VC route this time, because the problem I am going after feels VC size.

However I feel bootstrap or small F&F round gives you more flexibility when it comes to exits.

Congrats again!

Sure, but I imagine the terms of the acquisition doesn’t say you have to write it in this specific style.

I’m sure there’s a way to say the same thing without coming across as a bullshitter.

There is a theme in a lot of products (and open source projects) of plastering AI all over the description, even if it isn't really related to AI. The first thing I see on the CirrusLabs website is "Accelerate AI Adoption Without Compromising Safety". How am I supposed to know this is a CI product, and even if I figure that out, how am I supposed to know it is a genral CI tool and not something AI specific?

I don't understand why these sites just put a bunch of buzzwords instead of telling you what it actually is.

Not to sell Tart short (it is quite good), but it's "just" a wrapper around Virtualization.framework with a few extra pieces. This is the kind of thing that Codex driven by experts _should_ be able to build very easily.

Meh, Gehrig and Ruth are long since dead.

I mean, yes it is, but we all know what happens to these projects when their primary devs move on or get acquihired. Keybase, anyone?

If you are interested in yaml backed pipelines check out this open source tool I built for exactly this purpose:

https://github.com/smogili1/circuit

That sounds like a British phrase for pimping.

Every great journey must have an end. As far as I understand Cirrus CI was struggling due to Github Actions eating their market. Cirrus CI was in my opinion much better than Github Actions but it is hard to compete with a bundled solution.

You’ll be pleasantly surprised. Updates in the coming weeks.

> In the coming weeks, we will relicense all of our source-available tools, including Tart, Vetu and Orchard under a more permissive license. We have also stopped charging licensing fees for them.

I just don't think there's much upside to telling it how it is in the press release that gets buried after a week and everyone moves on. For better or worse.

You are checking the wrong Cirrus Labs.

Agreed. The benefit from not having anyone else or any partial (container) solutions in the computing chain is huge for secure isolation. Getting rid of the intermediary solves a universe of possible problems.

That said, I've been free-riding on tart because they've often surfaced issues I needed to address. Free riders like me are possibly the reason these companies can't make their own way.

Yes, but it’s also currently the best one. They have OCI compatible Mac VM images that are prebuilt. It’s quite good.

MIT or Apache2 or FreeBSD licenses would be preferable in my case, but GPLv2 or even AGPLv3 (if you have to) would work.

If you’re taking requests…

Why is telling it like it is not the rational default?

If this is a forgettable press release isn't it lower effort than coming up with this type of nonsense?

interesting that was what i thought this was, it keeps boggling my mind the sums being paid for what really could be built by experienced devs on their own teams

What was good about it? It looks pretty ordinary to me: https://cirrus-ci.org/features/

It gives a better docker based experience and it also has measurements of memory and cpu usage to help you dimension things quickly.

Short term: they don’t need devs to build it, it’s already built

Long term: they now have experienced dev(s?) to build their next products and features

Much better UX for viewing logs, more supported platforms. Github Actions in particular is also very unstable.