I ran the Webflow meetup for five years
And I wouldn’t build a new site on it today.
We don't look at websites the way we look at them in Figma. I wish more designers understood this.
I ran the San Diego Webflow Meetup for five years.
Until this year, Conversion Factory was a Webflow-exclusive shop.
I’ve defended that tool on more calls than I can count telling WordPress users that they were missing out on a better way to develop a marketing site.
So believe me when I say this one isn’t fun to write: I don’t think any new website should be built on Webflow. Or Framer. Or Squarespace, or WordPress (that last one isn't news, but worth repeating cuz WordPress sucks).
The sad part is that Webflow as a tool, compared to what it was last year, has gotten palpably better. But if it's gotten better and it was my recommended tool of choice for anyone looking for a marketing website, why has it become antiquated?
The short answer: Webflow is a middleman we don't need anymore.
Why we needed a middleman in the first place
Designers have always known what they wanted on the screen. What we didn’t have was a way to get it there without a front-end developer, and code was a language most of us never learned to speak.
So Webflow and Framer stood in the middle. A visual layer between the designer and the code, so you could move what you saw and let the tool translate it. For about a decade that was a genuinely great deal, and I built a large chunk of my career on it.
But at this point, the middle is causing more delays than progress.
Complex animation is a fight. Rapid ideation of different concepts is a fight. Pulling in data is a fight. A lot of what should be automatic is manual inside of Webflow, and the whole thing is slower and clunkier than it has any right to be in 2026. Ask for something outside its known phrasebook and you get the closest thing it knows how to design with a lot of refinement still needed.
Custom code is available to everybody now
Here’s what changed: coding agents are the highly skilled front-end developers designers were trying to avoid. They’re the hands and feet. The designer gets to be the head, which is the part we were always good at anyway.
You don’t have to avoid custom code anymore, you lean on it, because the thing that made it expensive (someone who could actually write it) is sitting in a window on your laptop and they are the most servile, obedient, tireless front-end developer you have ever worked with.
There is no way I could have built my personal site in Webflow.
At least, not without spending the better part of a year and tearing my hair out... from my beard 😂
Same goes for Prisma, RadioCult, Conversion Factory, Pricing Pages, and Parseur. Put any of them next to the best work our design team shipped on Webflow and it isn’t even a contest.
If you can dream it, you can ship it. The tool isn’t the ceiling anymore. You are.
The hard part is getting in the door
It'd be a lie to pretend this is frictionless. The hard part, the reason we needed Webflow at all, is still getting into the code.
Until recently that meant the terminal, and terminals suck.
Everything comes in monospace fonts, it's high contrast and jarring on the eyes. You can’t click on anything. Nothing renders the way a normal human expects it to.
But that’s changing fast. There’s a whole category of apps now that exist only to make working with agents pleasant: Conductor (what I use), Warp, Solo, the Claude desktop app. Once you’re in, it’s honestly not that hard.
But it is a different skill for most people who are used to doing things by hand to become acquainted with. You stop pushing pixels and start giving creative direction.
What creative direction looks like now
I record a Loom. I talk through what’s wrong, circle things, point at the spacing that’s off and the headline that’s too loud. Then my co-founder Corey’s watch-video skill lets the agent actually watch the video, and it goes and makes the changes.
When I need to get my hands on something, I copy the element as raw HTML, paste it into Paper, push it around until it’s right, and send it back to the agent. Paper’s canvas is HTML and CSS, so nothing gets lost on the way in or the way out.
Talk, point, tweak, send. Done.
Give the devil his due
Webflow’s CMS is still hard to beat, even compared to the next-in-line tool, Sanity. It’s easy to understand, easy to coach a client through, and the on-site editor is great. Can't sing enough praise about that aspect of their product.
It’s also the next thing to go in the "AI-can-handle-that" part of the equation. You don’t need an on-site editor when you can point at the page with something like Agentation, or record a Loom, and hand it to Claude Code. Same goes for Framer's CMS. (And the enterprise paperwork isn’t a reason either. Vercel has SOC 2 and ISO 27001 too.)
Webflow knows
In September Webflow announced Source, an orchestration layer that sits on top of custom-coded sites. That’s the hopeful part, and it’s why this isn’t a pure dunk.
But the thing they’ve been selling for years is a visual builder, and Source is a bet on a world that doesn’t need one. They’re in an interesting spot.
There’s no reason to put a new website on a visual builder anymore. Talk to an agent, design on a canvas, build it in code, and stop writing CSS by hand.
— Zach
PS. Still on Webflow? Reply and tell me what’s keeping you there. I read every one.
