Issue #40: Send It to Me
The metaframeworks scene (as the JavaScript tooling scene in general) changes constantly — new projects pop up, existing projects disappear or decline in adoption/popularity. I watch this process attentively for more than a year with the Metaframework Records newsletter and start to see patterns and trends.
The current issue is dedicated to new tools and nostalgic projects specifically, but even more than that to one simple and humble ask from me to you: if you spot a new metaframework or have some corresponding idea in mind — please ping me in response to any of the newsletter issues. I would love to share more about the versatile and rich ecosystem this publication (as well as the “Comparison” part of it and the Encyclopedia of Metaframeworks) is dedicated to. I’m not a moneybag and can hardly propose any reasonable bounty, but I can promise your name will be here, and from that moment you will become an indispensable part of this small open-source community. It ain’t much but consider it an invitation to truly rock together.
But if it’s not your piece of cake, you’re still welcome to the ride.
The Good
Brian Kim, the creator of Crank.js, has announced their new project — a metaframework based on this UI framework (but rather framework-agnostic in fact, or even “frameworkless”, if you will). It is called Shovel.js and its main idea and philosophy is using service workers and browsers’ native APIs (shimmed where missing) as a foundation for the universal rendering capabilities.
With Shovel, most code is just written in the fetch handler, regardless of rendering mode. The difference between the modes is just timing.
The project is at its early development stage now but the capabilities it provides are quite enticing, and the approaches to implementing them look innovative, straightforward, and explicit, as opposed to many modern tools. Of course, production-grade honing will make its sophistication have an impact, but for now it is interesting to see where further experimenting will lead the maintainers to.
The roadmap is ambitious: sessions, authentication, websockets, cron jobs, email; ultimately, I want Shovel to be maximally batteries included, complete with an admin interface like Django.
And while we welcome new innovative ideas and swear big corporations burying our favourite tools, Philippe Serhal does not sit back and instead perseveres in resurrecting the old nostalgic (my first personal website) Gatsby (which is kinda considered dead long ago for the Twitter inhabitants). The project got v5.16.0 release bringing in the React 19 and Node 24 stories and corresponding adjustments.
It is funny as the recently published State of JavaScript 2025 survey (the results of which I personally take with a huge bag of salt but still) indicates 1% growth of Gatsby’s usage and stable interest (as opposed to declining interest for other metaframeworks). And the neighbouring (and even less adequate) State of React 2025 highlights the whopping 4% interest growth… Philippe, it is your chance!
Jokes and pissing contests aside, the metaframeworks happiness seems to decline which brings some concerns to me as the maintainer of this newsletter, so I emphasize my initial call: let’s make this ecosystem shine, my friends, and find and build new awesome tools (or at least decrease the survey-indicated 14%-excessive complexity factor to 13.5% in 2026, it sounds doable).
The Bad
Another big trend (and arguably a problem) of 2025 and early 2026 is multiple venture-capital-driven acquisitions and mergers in the open-source ecosystem. Even my favourite reading resource of these days, DEV.to, got acquired which worries the hell out of me because we see how clearly the toxicity rate increases for some projects after venture capital comes to stakes. I always hope it all goes for the better (at least the maintainers benefit from that which is rad!) but I also dream that the open-source spirit of genuine exploration and innovation for the sake of the common good will live and thrive. At least 2.8% more than in this February.
Meanwhile, the one project that is not afraid of anything being the mainstream metaframeworks world OG, strives for the agentic future and going all in for the agentic experience with the hands of people with at least some reasonable developer experience and their feedback. FWIW this evolution looks more like parasitism to me, but we all have opinions. If you somewhat worried about the future of Next.js too though, you can check out some useful hints from Jacob Heric on that topic. But if you still decide to stick with the 22%-most-appreciated-metaframework-of-2025 for your blog in 2026, make sure you have patched the recently disclosed vulnerability in next-mdx-remote.
The Noteworthy
Talking about the TanStack, another person with strong opinions, Adam Rackis, has published the series of cool thoughtful deep-dive articles on single flight mutations (probably pivoting on the recent Fallout series popularity).
A “single flight mutation” is a fancy way of saying: mutate the data and update the UI with just one round trip to the network.
And when you hear “network performance” you can bravely bet the echo is “TanStack Start”, that’s for sure.
Another JS guru, Nicholas Zakas, announced the new major release of my favourite JS tool of all time, ESLint. It comes with multiple breaking changes improvements, including enhanced [new flat] configuration files traversal (especially for monorepos), better performance (eat this oxlint!), slicker JSX support, new rules, and a lot more.
With all the huge job the ESLint maintainers team had done lately, they’d definitely benefit from a vacation. And we started to see these open-source holidays happening more and more, which is cool. For instance, the recently emerged npmx project, providing a new way of consuming the information from npm registry, declared the official mandatory vacation recently which is… kinda cool! Speaking of that, I have no idea what one could do for the whole week without coding, it def sounds a bit lame, but I fancy figuring that out so I want to experiment and take a week off too, thus the next issue will come not in two weeks but in three instead, don’t be surprised (I hope I’ll have another surprise for you on my return though!).
Hence, I’ll see you next time in the spring, and let’s hope the innovations in the metaframework ecosystem and community proceed growing — like flowers in March. Or like new AI coding assistants in 2025. You do you.
🏖️