I appreciate you taking this seriously, and I want to say that clearly upfront.
You did not just describe a frustration, you built a prototype, you made it tangible, and you are clearly exploring multiple ways to improve what we already have. That matters. It is easy to point at shortcomings. It is much harder to turn that into something people can click through and react to. So thank you for putting real effort behind the idea, and for showing a willingness to meet the ecosystem where it is today instead of insisting everything must be rebuilt from scratch.
I am also genuinely excited by the direction you are taking with AI. I have worked with AI professionally, and yet it never crossed my mind to apply it to the extension index in a practical way. Seeing you use it for categorisation and summaries is one of those moments where you realise a new tool can unlock options that were simply not realistic before. Not as a replacement for real metadata or good READMEs, but as a bridge that helps people navigate and compare. That is a fresh angle, and it is valuable.
With that said, I want to add some context from my side so we can keep the conversation grounded in what is currently possible and why things look the way they do.
Yes, flarum.org is lacking in many ways. For a long time it has also been unclear to me what the main pain points were, and which changes would actually move the needle. I remember you raising frustrations in an earlier post. I tried to address what I could based on that, but I do not recall a follow-up that confirmed whether those changes helped or what still felt broken. That is not a complaint, it is simply why a concrete prototype like this helps a lot. It turns a vague sense of “this is not good enough” into specific areas we can evaluate.
Another important piece of context is responsibility and maintenance. Since the extension index existed on flarum.org, the site has effectively been my responsibility. The index was ported over from Extiverse, so what exists today is a combined codebase. Over the years, core developers have shown interest in contributing. People were given access and a tour of how things work, but participation never really followed through. In practice that left the full weight of maintaining and evolving flarum.org on me.
The extension index also evolved into more than an index. It became part of how open source extension developers try to become sustainable, including premium subscriptions. That changed the stakes. It raised the bar for stability, policy decisions, and predictable behaviour. At the same time, I have been in burnout, and balancing the project with personal life and my own career has made it extremely difficult to consistently invest energy into flarum.org.
This is also why I want to address the common perception that only well known developers show up. There is no intentional promotion logic that boosts one developer over another.
There is some logic in the background, but it is limited and primarily meant to reduce confusion.
If an extension is abandoned and the author has set a replacement on Packagist, flarum.org can detect that and will hide the abandoned extension from default browsing. In general, abandoned extensions are mostly hidden in listings, but they are not removed. You can still access an extension page directly if you have the URL.
The other major mechanism is the compatibility check with core. This is probably the main reason an extension might not show in certain contexts. The goal is to avoid advertising extensions that clearly do not match the selected core version. That logic did misbehave in the past, but as far as I can tell it has been working reliably for a while now, including throughout the 2.0 betas, unless I am missing a current edge case.
Aside from those two things, there is nothing that blocks extensions from showing up. If the list feels dominated by certain names, it is likely a side effect of the signals we do display, especially download counts, and the reality that a small number of packages account for a large share of installs. That is exactly where your proposal is interesting, because it tries to introduce better discovery pathways and more meaningful context than raw popularity.
On reviews and comments, I also want to share some history. I did work on a review or comment style system in the past. The core team discussion at the time was that it would move interaction away from Discuss. On top of that, when I tested a simple rating feature, it quickly devolved into fake ratings and low quality submissions, often complaints rather than useful evaluation. That does not mean it is impossible, but it does mean any review system needs careful design and moderation thinking, otherwise it becomes noise and workload instead of signal.
The uncomfortable part is the future. The future of flarum.org is currently uncertain. I would love to commit to implementing a number of the ideas you show here. I am just not sure right now whether it is necessary, or whether I am capable energy wise, to promise that kind of work and follow through properly.
What I do want, is to avoid this ending with applause and no progress. A call is scheduled this week to talk things over with Gregor, your proposal is definitely going to be part of the discussion.
Thank you so much for all you do @huseyinfiliz, I hope you understand by now that you already deserved my appreciation and gratitude ten times over.
