I wanted to let you know that I think Codefloe is a fantastic service and its what a lot of people are looking for right now.
However, I did not think that the first two times I came into the landing page (after mentions on Hacker News). The landing page content is ok but its not engaging, and the style is just themed Forgejo and so it comes across initially as a random public instance.
In contrast, when I land on your docs home page - it immediately comes across that you are taking a very serious approach to how you manage this platform and what your goals are. And the style there - while somewhat generic for a landing page - is still much better than the actual landing page.
I think you’ve got a window right now, where a lot of people are looking for alternatives to GitHub and Gitlab.com. I would re-do your landing page - if nothing else start with a new version from docs landing page on docs.codefloe.com, but focus on the things most immediately relevant to developers looking for a new forge:
personal projects & private repos are allowed
No ideological filters on suitable projects
commercial / business projects are fine
excellent CI/CD service with free tier (is there a paid tier yet? for a business they’d prefer to be paying you in many cases)
Please understand this feedback comes from a true intent to want to see you succeed and take advantage of an opportunity that may not come around again.
@jeremyjh Thanks for sharing your thoughts! Such feedback is very valuable to us, especially when it comes from people who like the platform and want to see it getting even better
Yes, when we started out, we (at least) wanted to have a custom landing page but don’t “complicate” matters too much, so we went with the default template. Agreed that this might not be the overall best layout to promote a project which is quite ambitious in their values and goes further than a “random” public instance that has everything at FJ defaults.
The docs landing page comes from the default Astro Starlight style which comes with a landing page by default. It can be turned off though. Ideally the docs don’t have a landing page at all and this part is handled by entry page at codefloe.com directly. What do you think?
We absolutely agree. Given we don’t have a profit goal per se (see our financial approach for details), we won’t do any ads or similar to promote the site - at least not before all infra costs are covered by donations. Hence it is a bit harder to get the message out to people who might be searching for something like this. Within the existing Codeberg/FJ community, we got more backpush than support after our announcement and in general (presumably due to being viewed as “the unwanted child that isn’t Codeberg” and now even “welcomes” AI projects as both CB and FJ are strictly against it). But this is another topic
There isn’t a classical tier you can directly subscribe to as in cloud offerings today. We are not sure it will ever be but we might not rule it out either. We currently offer to provide more storage (and at some point also CICD - usage limits are still missing for this area) for larger orgs who want/need these. Today the approach is to reach out to use and refer the “donation” contribution as a meaningful asset for over-average consumption use.
PS: we are always open to other contributing to the platform also on a technical level. If this is ever of interest to you or others reading here, please reach out and/or post in Join the CodeFloe infrastructure team! .
I think you might want two still, and docs are fine the way they are for docs. For the main landing page, I wouldn’t include all the content from the docs landing - I’d focus on content that directly speaks to why people might be looking for this service. So the CI/CD & usage content you’d definitely want, but otherwise I’d focus on the other things I mentioned.
Well sure, they already have Codeberg, but they are a tiny, tiny fraction of the whole audience you could appeal to. This is what is happening: every few days, GitHub has another major outage or degraded service. A thread will stay on the front page of Hacker News for most of that day and get tens of thousands of views. The thread is always largely focused on what the alternatives are.
For a long while, Codeberg was the heir apparent. That is no longer true. Your service gets mentioned in almost every thread like this, and people come here and look at your landing page. That is the opportunity. GitHub will stabilize and return to the status quo. Tangled will mature and someday offer private repos. So thats why I think you have a narrow window.
And yes - I know you aren’t strictly profit-seeking. But you want to get your user base large enough to have good utilization of the hardware already allocated and donations that cover costs, and get more people involved so it becomes a real community.
Yes, I meant paid tier for the storage & CI/CD limits. I would prioritize CI/CD limits because it could become a real issue if you have a large influx. Demand for that is virtually infinite and it doesn’t necessarily benefit the community to have lots of heavy CI/CD users (that could be all they are using you for).
One other thing I forgot to mention - if you haven’t done it recently, open codeberg.org in a private browser window, so you see their landing page. Its actually really good!
Yes, that’s a good reasoning. In general, we’re also happy if other alternatives emerge so people actually have a choice. For a long time, there was only GL and it was close to GH while having a very opinionated UI and other quirks which devs don’t like. For a long time, there was no “dev oriented” FJ-based instance that focused on performance and dev workflows - this is what we want to offer with CodeFloe. Ideally in a federated ecosystem long-term, but we’ll have to see if this ever becomes reality.
On point!
Yes, this is on our near-future roadmap. For now, we wanted to make the dev experience first class (top notch DB, fast text and code search, fast CI). For the time being we can take it without any limits on that sector but we are aware that this won’t work out long-term.
Hello, I just want to express my support to Jeremy’s idea. It’s been some time that I’m using Codefloe fairly regularly and it wasn’t that easy to find out about its existence and not that obvious at the first sight how it differs from Codeberg for instance (but also other public FJ instances), i.e. being more open to private and commercial repositories back then and now also to AI. Also, an important differentiator for me is that Codefloe is faster than Codeberg which is noticeable just with a naked eye during git interactions.
And honestly, once I’ve read the documentation regarding the infrastructure solution and had a look at the infra repo, I can only say how much I appreciate your proactive, smart and data safe approach. Which would also be another point important to many EU/EFTA devs who don’t want to maintain their own instance, but at the same time want to use a reliable and always up-to-date product.
I hope I could contribute to the project in the future and also become one of your commercial customers.
Great job!
Yes exactly - I alluded to this but thank you for calling it out more specifically. Because I did exactly the same thing - saw the docs page, followed links to the IAC repo and found the hosting & infrastructure docs etc. I doubt there is currently another public instance hosted this well, with HA database and all IAC etc. So, I think this is why people mention codefloe.com on Hacker News, but not everyone will dig that deep to discover it.
Also, while I share the goal of a federation of smaller instances (and I have a lot more about to say about this in another topic), I think @codefloe is by far in the best position to lead the charge there. It can demonstrate that people can use a platform like this and it works well. Basically proving product/market fit. Also the transparent ops and providing all the docs it helps others see how to follow the example.
We’re super happy to take hints on how we can improve that! Promoting ourselves initially was (more likely is) (very) hard as people don’t like reading about “another Codeberg” on Mastodon and alike which claimed objective technical and operational shortcomings in Codeberg as the main reason for the creation. Meanwhile, there are more philosophical differences with CB as they have done a strong move towards an association taking (mainly) political stances in all areas whereas the technical part feels like a secondary importance. For us the joy is in/about the (technical) platform part. Anyways…
The best promotion is always coming from happy users talking about their experience in external places - and you’re right, once interested people got a pointer, the ball is back on our side and we need to place the important bits where people might expect it. And we surely can do better!
This was very important to us from day zero, especially because Forgejo has the potential to be very snappy and we found it really sad to see that the main impression the world was seeing is the slow speed CB is/was delivering. And no, it’s not necessarily a question of “load”: the very same discussion is going on for many years and back then CB was also slow, at some points even slower than today, while the load was roughly 1/20 or even less. If there’s a bottleneck in the infra setup, it causes problems even for small/medium sized instances. On the opposite, if done right, one doesn’t “feel” it even if theres X times more “load”.
Such messages really make our day, thanks for posting this!
And it’s exactly this what we want to offer: everyone talks about IaC, pipelines and automation. Yet almost nobody builds a platform in the open and shares the config and underlying hardware, even though these are essential building blocks.
It motivates us to see our configs protecting the instance while at the same time being able to server a rocking-fast platform on well-configured and modern hardware and showcasing that we can (hopefully) many thousand active users with it.
The possible next “promotion” step would likely be some big tech influencers picking us up and showcasing the instance in their community. Or some blog posts from people with a large audience. Yet, it would feel wrong to us to “ask” them proactively or even do so for money. Sure, it would effectively “just” be a investment into platform promotion but we hope/believe that this might happen eventually if we just keep on improving the platform and offer something that has added value for people
PS: we’re currently rebuilding the main landing page and hope that this will contribute to such small continuous improvements - thanks for mentioning again!
