Pico CSS is no longer maintained
github.com
github.com
Why? Why? Why? I hate how everyone seems convinced they need to adopt the latest and greatest browser features just to stay relevant. I hate oklch and I hate Tailwind popularized it. I don't care how scientifically accurate it is at representing human color perception. It's a PITA to understand a new type of color system, and learn to tweak it, especially when HSL/RGB just worked fine.
This just seems to me like (1) not wanting to learn new things, (2) a logical fallacy because the conclusion that those changes are to "stay relevant" and not to make the life of any future mantainer easier, and (3) fear of the new (https://en.wikipedia.org/wiki/Neophobia)
There's not "120 degrees of red, 120 degrees of green, 120 degrees of blue", so blue is not 240. Actual bright blue is hue 264. The "0-255 RGB" bright blue is 0.452 lightness, and max chroma on a really good monitor is something like 0.33. The lightness mostly makes sense, right? 0.5 is halfway between totally dark and completely washed out.
"Ok, so chroma tops at at 0.33" not so fast. You can crank up the chroma to 0.40 and still get either magenta (where lightness is around 0.74) or green (lightness around 0.78). So now you need to know a valid, nonlinear lightness range for each band of colors, because midpoint blue is different than midpoint red which is different from midpoint green.
From greyscale luminosity calculations and knowing about subpixels and bayer filters, I would've been able to guess that green is the color with the highest chroma, but then there's bright magenta, neck-and-neck. Oh. Until I move the lightness slider down on the web color picker, and then everything (especially peaks) in the chroma channel shifts around dramatically... So much for intuition.
While OKLCH makes it really convenient to select colors with properties you like along a gradient, the nonlinearity means either memorizing a bunch of base colors, using a color picker, or converting from a system that is "actually" intuitive. I can't really look at a trio of OKLCH numbers and predict the output color like I'd be able to with RGB, although this is mindful that RGB is also not linear (for the exact opposite reason as OKLCH) and I can only do that from decades of experience.
I'll concede that magenta and bright green and bold blue don't really have a place on the world wide web other than in a company logo, so these are outliers. Still, the OKLCH format is not what I'd call intuitive to a total beginner.
(Also, I'd very much not like to implement OKLCH calculations on a microcontroller until lighting companies start selling OKLCH LEDs.)
What justification do you have to adopt and replace this with existing solutions with such urgency? Have you researched its compatibility? Are you providing any fallback?
This just seem to like 1) you are not applying your mind why you want this, 2) faulty premise that it makes it easier, because in reality your software never reaches a point of maturity, where its just "done", and 3) shiny object syndrome (https://en.wikipedia.org/wiki/Shiny_object_syndrome)
After I read the same I thought, ok, that's what _they_ wanted to do.
Me? I'd love to see what could be rewritten to be 100% future proofed for as many years as possible. But see what that did? I have to share, and see if anyone else already thought of this (I doubt I'm the first ever to have this thought), and maybe, I'll be lucky enough to benefit from someone exploring it already, before I decide to put on a jedi robe and wave my hands to use the force to move AI around.
Alas, in terms of the software people use to mock up designs, very few desktop apps support it. HSL is, for now, the more practical option.
Take bright yellows as one of the worst examples. Basically the only thing you can do with them is reduce the chroma: any other change will take you out of gamut. As the most extreme case, #ff0 is approximately oklch(0.968 0.211 109.77), and the following minuscule modifications will take it out of gamut: L + 0.0002, L − 0.0001, C + 0.0002, H − 0.009 or H + 0.025.
This is why the creator massaged them into a cube/cylinder: https://bottosson.github.io/posts/colorpicker/, Okhsv and Okhsl. Those are much closer to being uniformly superior to sRGB.
Presumably we can deprecate the broken shit if we can get people to buy into software upgrades.
I see how position-anchor and stuff are coll for complex layouts, but they have poor support on low-power mobile browsers, and, frankly, are not that useful for a minimalistic CSS framework. If you want complicated layouts with non-obvious anchoring rules, you're not the target audience.
Cascade layers are very useful for complex designs with many independently added stylesheets. They would be great to have. OTOH again, if you need this level of complexity, you're likely not the target audience of a pico-sized CSS framework.
I would say that a minimalist framework should make judicious use of CSS3 features, and not try to be excessively clever. In reliable engineering, being smart often means avoiding being clever.
- Nimble: https://leftium.github.io/nimble.css/demo/pico.html
- Sumi.css: https://vladocar.github.io/Sumi.css/docs.html
- Blades: https://github.com/anyblades/blades
- Oat: https://oat.ink/
From a quick look at the second horizontal strip that starts "what is this?", and without wasting time doing too much digging, it appears to be ultimately caused by using a grid layout for a heading followed by a few paragraphs, all of which fill the available (constrained) width, which amounts to one of the absolute simplest layouts possible in CSS - block. Last time I saw something similar, albeit with flexbox not grid, it was from a junior developer who appeared to have only learned flexbox and nothing else.
As a quick semi-fix, just disabling grid-template-columns on .base-layout is enough to fix almost everything except the constrained width.
The layout approach is based on this Kevin Powell video: https://m.youtube.com/watch?v=c13gpBrnGEw
It uses CSS Grid to create a centered, width-constrained content area while allowing individual elements to break out of it or span the full viewport width. This provides a flexible, consistent layout without the need for additional wrapper elements in the HTML.
CSS nesting has had full browser support for more than 3 years now, I recommend updating. I'll also add fallback for older browsers, but the Grid layout itself isn't the issue.
Yes, that's right. I thought I'd try it in that older browser that I still happen to have installed, as the user above didn't immediately let us know what that mobile browser was. It may or may not help by at least giving you a browser version you can use to test. As I mentioned somewhere else, it makes sense to use some sort of older browser as a proxy for older browsers in general, when testing.
Developers are always looking for new ways to do what we've always done, often using more technically complicated ways to recreate the simplest things. It's fun from the puzzle-solving point of view, but your choice of earliest browsers you want to support depends on your user base.
How's your ten second evaluation of it?
It's one of the reasons I'm planning on moving stuff over to Pico.
Or now, I guess, just writing some CSS from scratch, which I hate.
Ios 13.6 in my case, which was released just 6 yrs ago, 2020. For some context, flexbox gap was added in ios14.1, released in 2021. So iOS 13.6 was still usable atleast couple of years beyond that. So such frameworks utterly failing on such browsers is hilarious, given that they market themselves as minimal.
People keep forgetting that css3 was announced in 1999. And adoption roughly started from 2010. The css3 of that time is not the same as css3 you get to see in latest browsers. These minimal css frameworks are really useless if you don’t know what css3 features you are writing and who you are targeting.
And, I Haven’t even tested those in less capable browsers like dillo or netsurf, but I doubt they will even work there.
If you start looking into browser market share, you will quickly realise that there are millions of worldwide users who can’t use the web normally.
Minimal is an entirely different property to backwards compatibility.
motherfuckingwebsite.com is more minimal and usable than these frameworks.
No you do not. You just saying it’s got very little code and/or very few features.
Support for older browsers might not be one of those features. Which means this minimal frameworks isn’t suited for the kind of projects where that kind of compatibility is a concern.
If you look outside of CSS (and particularly web development as a whole) you’ll find a plethora of minimal frameworks that intentionally make compatibility trade offs.
> Which in this case is clearly [does not work].
That’s a gross exaggeration and you know it.
For a long time we were forced to use the ugliest hacks imaginable to get to the nice (responsive, design-minimalist(!)) layouts we wanted, because CSS layout tools were utter archaic bs, for years on end, because browsers iterated so slowly.
Then things started speeding up, and we're getting to make better websites (and apps!) that are way easier to design (with the correct design theory equivalents), implenent (with less code) and easier to maintain (i.e., understand). And you're not missing out if you just keep your browser up to date.
I know it sucks if you're on an unsupported device that doesn't get any updates, fair.
But let's face it: 99% of the population is better off security-wise when we're not enabling them to run unpatched devices forever.
Much of the issues of layout were already solved more than 12 yrs back, ie when css3 became mainstream.
The issue is devs are unaware about different levels of css3 or just plain ignorant about it.
Are you aware about the different levels of css3? Are you even aware there are levels to css3? Why are these frameworks not listing them and simply saying that they comply with web standards? Which specific web standards do they comply?
Any new css3 feature should have a fallback if they say they comply to web standards, which these frameworks don’t add.
Basically you want to adopt latest w3c guideline but completely ignore wcag guidelines on accessibility - https://www.thewcag.com/principles/robust
Why should some population forced to suffer because of your ignorance. I am sorry if I am blunt, but your comment about polarising users is rude.
Which comes back to my main point, Devs are simply not educating themselves about the fundamentals of web design.
There are older browsers which still get security patches?
ios updates unfortunately ties everything to a single batch. Unlike android. I have an older blackberry, that still gets latest app updates.
What is even hilarious is ads render with pixel perfection on older devices. So for some reason, ad networks care about browser compatibility than devs putting the same ads on their incompatible website.
And yes, it makes sense to use one particular old browser as a general proxy for older browsers in general, so people shouldn't get hung up on the particular version of Safari you happened to mention.
I don’t care about what’s trendy, but using outdated, insecure browsers to surf the web is just plain stupid. Even more, expecting developers to waste time working around all the bugs in those old browsers is just insane. I do the opposite: as soon as a browser reaches EOL I drop the support. Don't encourage lazy admins and bad habits.
> expecting developers to waste time working around all the bugs in those old browsers is just insane
What bugs? The ancestor posts are talking about frameworks immediately using features that were just added to browsers - i.e. more recently than those older versions. Nothing at all to do with bugs. Can you help us to see what you've misunderstood here? Perhaps not dismissing users as "stupid", developers as "insane", and admins (?) as "lazy", would be a good starting point.
Seems to be a bad decision from the beginning.
> And developers really should know better than to fall for the marketing nonsense that anything more than 5 minutes old is "insecure".
Where have you been the last decades? Pick an arbitrary browser and go through their CVE list, e.g.: https://app.opencve.io/cve/?product=safari&vendor=apple
> What bugs?
I guess at that point I have to assume you have no idea about front-end development. Otherwise you would know what a pain in the ass old browsers with their inconsistencies are (e.g. remember the struggle with the early flexbox implementations?).
I don't disagree! But at the same time, part of our job as web developers involves acknowledging the reality of what our users actually do.
What do early flexbox implementations (something like 12-15 years ago?) have to do with anything? We're not talking about browsers as old as that! They weren't even bugs, they were early implementations of early versions of the standard, so who exactly is the one with no idea here? That level of browser inconsistencies is long gone.
Those of us with years of experience of building user-centric front-ends are not the ones who jump to use new browser features too quickly.
I won't be continuing this conversation, due to the general arrogance and condescension displayed in your posts - following "stupid", "insane" and "lazy" with "bad decision", "where have you been" and "you have no idea".
It shows that a) new technologies may add value and also that browsers constantly fix bugs due to complex or underspecified standards. It was just an example, go through the issues lists and you find a lot more.
> not the ones who jump to use new browser features too quickly
That's reasonable but there is a point where support just should be dropped. Don't support IE 5 for ages just because someone is too scared to update. Let them fix their technical debt. Encouraging people to access the internet with insecure software is not acceptable. Btw. that was the argument you completely ignored in your reply.
Thank you Lucas Larroche for building Pico CSS and sharing it with us.
> Today, AI can generate lightweight, standalone, accessible HTML with just the CSS it needs, so a Minimal CSS Framework for Semantic HTML matters less than it used to.
If your answer is "use AI", then why not use AI to not need Pico. This could be maintained for those of us avoiding AI, of course, but we are fewer and further between with each passing day, and likely rolled our own instead of using someone else's framework anyway, so he probably didn't see the point and wanted to spend the time elsewhere.
> brand recognition that shouldn't be throw away
How useful do you think this brand recognition would be in his life? Enough for it to be worth maintaining a project he had presumably moved away from using for his own work & play?
If it is so valuable then maybe someone will make a deal to take over maintainance. Though TBH if this were mine I'd probably prefer to keep the project mine-but-archived than let someone else do what they will with the "brand" thereby diluting what value it might have to me personally. It can just be forked instead, or people can move into one of the already existing forks.
Thank you to the developer and maintainer
For what it’s worth I had to force Claude early in to use pico and not ever generate react or ornate tailwind.
Pico is extraordinary for standalone tools where ai can’t let complexity build out of control.
Maybe there’s other options but pico is what got it done, and vanilla LLMs has a hard time keeping something this simple and elegant and flexible.
This argument is basically saying, "Why build anything if AI can do something similar?" Personally, I think there is a great deal of value in building things to a well thought out specification. I don't like this attitude of cognitive surrender.
We’re actively maintaining Blades CSS on GitHub and Discord, and would love Community feedback there.
We also set up a dedicated page outlining key details: https://blades.dev/css/pico/
Ever since Coding Assistant LLMs became pretty good (2025-Q4 - 2026-Q1), I feel that it makes very little sense to continue to use pre-defined CSS Frameworks that 'has everything.' Of course, it is still OK if the general team decides on one, and they are OK (such as the Tailwinds, MUI of the frontend world.)
I have used Foundation[1] and Bourbon[2] extensively. Mostly because I can break it down and take just what the project needs and keep it within a low CSS budget for any project.
With that in mind, in the 10+ projects that started for some of the biggest brands of the world, I just defined the CSS framework, design systems, design token, with room for extension as the project grows. Claude Code is my tool of choice for now and I just let it run overnight, tweak as the project scaffolds and finally takes about less than a week to have a complete CSS framework in just plain CSS, and JavaScript where it needs to talk to the broader frontend.
Guide for replacing its features : https://thoughtbot.com/blog/you-might-not-need-bourbon
I mean with AI you can possibly more easily make decisions on the motor oil and suspension choices of the car you are building... but there may still be people who only want to get an pre-designed car, and only make calls on trim and color maybe.
It was a blast, super simple markup, mostly good defaults, I really enjoyed it.
I hope the best for them, and I hope some fork may get some traction, as Pico was my goto framework for small stuff one needs to do every now and then, and I really don't like the tailwind approach.
I might need a CSS solution that is less React-heavy, one that provides basic components and a user-friendly style system; I just need to include a CSS file (possibly a helper JS file) and then I can make style adjustments in my own application CSS.
Do you have any recommendations?
Do people actually use this? In particular, before LLMs ate the world, I mean. When I did frontend, it was hex, or if you felt fancy, `rgba(...)` (though at that point named colours had mostly fallen out of style).
Is using these less-intuitive colour encodings made more practical through the use of CSS variables?
RGB is fine for specifying a color. I mean, there are issues with RGB colorspaces and whatnot, but if you need to pick the background of a page or the color of text, there's no real reason not to use RGB.
The problem with RGB is that it is really bad at color interpolation. You can't take the average of the red, green, and blue components of two colors and expect a usable color in the middle to come out. Instead, you need to move to a perceptual color space. CIELAB is the old perceptual color space of choice; Oklab is the new perceptual color space of choice.
Color interpolation is needed for things like gradients or information visualization, where you need to say things like "20% of the way between color A and color B".
Yes. OKLCH is actually MORE intuitive when creating shades because it's based around human perception. It also supports wide color spaces like P3, where rgba is, IIRC, bound to sRGB.
The utility of these color spaces and color functions isn't lost just because LLM. Color systems haven't gone anywhere.
Would highly recommend reading up on OKLCH if you haven't already. A lot of people put a lot of effort into promoting this and bringing it to baseline..
So in my case where color shifting is happening, Oklab actually helps. I don’t know how it’s better than HSL for simple UI elements unless maybe you can get physically brighter values that pop out more on certain screens.
This started way back working with my long-term business partner and he built a pretty good tool to help visualize and tinker with it.
Website https://colorcolor.in
https://evilmartians.com/chronicles/oklch-in-css-why-quit-rg...
I find it more intuitive: when I want to mix colors, or algorithmically choose colors, only oklch works to give decent results.
Maybe you want to severely darken/lighten a color on a page when the user hovers over it - doing it in RGB or HSL results in a color that does not go with the existing colors. Doing it in OKLCH does.
Do people actually use this?
I love OKLCH. Awkward to say, beautiful to look at.
To be honest: I haven't really looked at the CSS that agents have written for me. I'm wondering if CSS frameworks still can play a role in that world as guidelines to get more consistent and maintenable code. On the other hand: agent does it, agent has to fix it.
Still not totally sure where all of this is going.
It's build-time, Rust/Oxc, type-safe, zero-runtime CSS-in-JS, with per-route CSS splitting.
The problem is that the explanation given here makes only little sense. I am not going to rely or depend on AI slop, so there is now one less option for me.
I think humans need to stop believing in AI slop as the new religion for everyone. Some people have a complete anti-AI policy and pull it through, too.
Does anyone know a similar project doing this?
Already plain modern CSS instead of Sass, has Popover API, and CSS layers for individual components, OKLCH colors are on the way.
Drop-in compatible with Pico CSS (via underlying fork)