I love Cloudflare. I use it for all of my hosting. I give its blog a heavy bias in my newsletter. I let its runtime decide which frameworks I use. It’s the only reason I have no Angular SSR projects despite wanting to try it out.
Cloudflare is the reason I am rewriting this blog with Hono. It’s entirely native to Cloudflare’s edge runtime without adapters.
But wtf is vinext? The announcement post says it’s a vibe-coded rebuild of NextJS that deploys to Cloudflare Workers.
We’ve seen OpenNext, we’ve seen Blitz.js. The wrappers, the abstractions, the fixes, the forks. All of them misunderstand the key philosophy of Next. The first thing you must consider before choosing it. Framework-defined infrastructure.
NextJS is a Vercel framework built by Vercel, for Vercel. Just as Amplify is an AWS framework. Just as Hydrogen is a Shopify framework. Just as Firebase is a Google Cloud framework.
But you don’t see the same wild attempts at making these something they’re not. I understand that these things are good. They tempt you into using the platform. When you don’t want to use the platform, you want the tool without the platform. But they are built to pull you into the platform.
What doesn’t make sense to me is wrapping, adapting, forking, and now vibing copies just to take these frameworks out of their natural habitat. It’s a lot of work to no good end.
If you’re using Cloudflare, you’re dealing with light workers and CDN. It does exceptionally well with Vita SPA, Astro, and open React SSR frameworks like TanStack.
If you’re using a VPS on another provider, you get the benefit of using more than just JavaScript. Spring, ASP.NET, Ruby on Rails, Django, Go, Rust, it’s all open. Dozens of solutions that are just as good, if not better than NextJS.
The only time choosing Next makes sense is to take advantage of Framework-defined infrastructure.
Now, instead of forking, we’re vibe coding copies. But for all intents and purposes it’s still a fork.
What got my undies in a special sort of bunch is the surrounding dialog. Over on GitHub someone asked if this project is serious. A legitimate question any reasonable person would have. Immediately asking if this is meant to improve on NextJS API’s, using it as a starting point. To be something unique. (Spoiler: The answer to that is no.)
Seeing someone thinking the same as I am, I added another question. There are still issues with Angular SSR and TanStack Start on Cloudflare. Why do we need a new framework when these can be improved?
This comment was marked as spam. Then from the viber in chief we get this comment.
I can say that with some authority. Yes, I’m the one who wrote most of this project, but I’m also the director in charge of the entire Cloudflare Workers org, almost 80+ people at this point. I’m not just an IC engineer who has to find time to justify continuing to work on this. I can literally put people on it, and I’ve already been talking to the team about how to do exactly that.
Listen, everyone has their space. Everyone is allowed to moderate their online spaces however they want. But we have to talk about how funny this paragraph is. I don’t know how to respond to this without being mean. I’m not Jon Stewart. But “I’m the director in charge, I can literally put people on it.” Well millennials, we made it. We’re old and we still talk like this. Millennial core at its finest.
As far as Next.js’s APIs: the current goal is to maintain parity with Next. I’ve definitely heard from people that there are things about Next they don’t like, and I appreciate that perspective. But right now, diverging significantly from Next’s API surface isn’t really in scope. Obviously, we already me an exception with TPR, which is an alternative take on pre-rendering, but the bulk of the Next.js APIs we’re trying to match 1:1. The reason is straightforward: we’re trying to give customers who are having a hard time hosting Next on Cloudflare a better time.
Feature parity is a classic fork killer problem. Especially when the fork is not maintained by the internal team based on the same conversations. And it’s never worth the maintenance. But considering the project is just vibed against forked tests, it’s not built on a stable, reliable foundation to begin with. Cloudflare, please consider the depth and accuracy of the Microslop meme.
Being labeled as spam by a millennial core director of literally 80+ people at this point hurt by butt a lot. So I responded with an angrier attempt at getting my point on this thread. Not speaking to Mr. Director himself but anyone naive enough to place their bets on a vibe-fork.
I should make it clear at this point that I do not consider myself to be in this little JavaScript web framework club. Which seems to be shrinking by the year. The drama and strange behavior are worse than that of YouTube influencers.
Here comes Tanner Linsley in an odd defense:
I’m flattered to see the FanStack out in full force…
Excuse me? Where? This is the type of ego that really puts me in a bad headspace. And ironically faintly representative of how out of touch and condescending a lot of tech has become.
I like TanStack almost as much as I like Cloudflare. But I think Start is mediocre. Good enough, but mediocre. How’s that for “FanStack”?
It’s interesting that Tanner is so deep in the JavaScript fandom that when he sees mere reference to a TanStack library he pictures a “fan” with the manic energy of a young woman at a Michael Jackson concert.
Getting comments hidden, trying again, being put into a creepy little box, and then seeing the conversation locked is the perfect condition for an ADHD+C-PTSD brain to ruminate at 2 AM. And it’s also a way to miss feedback.
We’re going to see where this project goes, what kind of feedback we get, and figure out where to take it from there. Hopefully that answers some of your questions.
The last thing I’ll say is this: if I could create a copy of Next in a week using only AI, you could probably build a framework that mirrors Next in the ways you like and solves the problems you’re describing in even less time. Honestly you should just try. The hard parts of framework development has gotten a lot easier.
NextJS is a great framework on Vercel and a terrible one everywhere else. The answer isn’t a vibe-fork.
I would love to see Steve Faulkner put his money where his mouth is and try creating his own. I would love to see Cloudflare actually build their own framework designed from the ground up for the platform. Not take from someone else and maintain feature parity. They could have easily created another OpenNext for Cloudflare. It’s not as if hosting Next on Cloudflare was impossible before. Instead, he chose the more complicated but more shiny and new path, a vibe-fork.
Real innovation. A client-centric framework built for Workers and the Cloudflare platform would be great. Hono makes for a good server. Astro is extraordinary for static sites. But there is nothing Cloudflare-centric like React.
Is it necessary to create a new framework? No. React already works. React frameworks (I shall not name any for fear of being talked past and down to as a “fan”) already exist.
So what niche does vinext actually fulfill? Next syntax on workers? Already exists. With a lack of open conversation and only “you’re doing great” gold sticker comments allowed, we have to let our imaginations fly for a moment.
It seems to me we have someone who sees a lot of people trying to put NextJS where it doesn’t belong. This is common, I’ve seen it and dealt with it. It’s been a problem since NextJS was added to the React docs. Before that it was a lowly niche framework for bloggers. When working with a Sitecore migration, speaking to the Sitecore team, I realized they were forced into it even though Sitecore already had a server side framework built on top of React. Including NextJS for no good reason only caused more problems and never really solved any. They seemed baffled why they were being made to make it work to begin with.
So this person who sees these problems decided to try to vibe code a solution for over $1,000 over a weekend. (Boasting as if that were a small amount of money, I might add.) They then jumped to the announcement Monday morning when they saw that ever so trustworthy AI spit out passing tests.
It’s exciting. I get it. I understand. But I’m nobody’s fanboy. And I think this is frankly a little delusional.
Will hobbyists use it? Yes. Will people complain that it’s broken? Yes. Will a few of these literally 80+ people be made to fix it? Yes.
And boy there we have the image of our industry in 2026. A weekend vibe-fork floating its way to the company blog and maintained with extra overhead over existing frameworks or even a new one.
I’m tired. The JavaScript ecosystem has always been an oddball bunch inventing problems. But at least for a few periods in time it was fun. Today we see less innovation, less systems thinking, less problem-solving. I look around and see people at the slot machines of AI turning silly weekend thoughts into pointless slop. I miss developers.