Webflow isn't dead! MCP 2.0

All right, folks. I have been hearing a lot of talk about how AI has killed Webflow, how Sanity is the new visual editor, and vibe coding is the future. Now, I am here as a Webflow professional partner. I've been working in Webflow for 5 or 6 years now. I'm here to say I don't quite agree with those things. I think Webflow still has some juice in the tank, and I think it's still a good choice. So, here are a couple of arguments that I have in favor of Webflow here in 2026. So, you might say I'm biased because I've been working in Webflow for so long, but my opinions are coming from working directly with clients and seeing how they are reacting to new advances in AI technology, how they want to use AI, how their bosses want them to use AI, and how I use AI to get along with all those things.

So, I don't think there's one-size-fits-all, of course, but I still think Webflow is a great choice. For a clickbait title, I almost went with the Webflow MCP is good now, but I chose Webflow is not dead instead. My first point here is really number one, Webflow has not done anything wrong. I was on a webinar last week listening to a webinar between Jesse from Finsweet and Mason Poe from Edgar Allan. So, those are some of the smartest people in the industry, some of the biggest Webflow agencies, and the longest-running Webflow agencies, and they both agreed that Webflow is still a strong offering, and they haven't done anything wrong.

That was Jesse's point. They haven't done anything that we would consider a grave mistake or a big blunder in terms of direction and strategy. That for me right there is a strong reason to stick with Webflow. We really actually don't want a platform that is trying to pivot dozens of times every quarter, trying to keep up with every new advance, and that really doesn't scream security and solidity from a platform. I know Vlad Magdalin isn't really front-running the company anymore as he was, but once upon a time on stage, he came to us freelancers and the community and said they at Webflow understand how impactful their platform is and how many people have been building and build their lives and their careers on Webflow.

So, I think the company still has that same spirit even though Vlad isn't the frontman day-to-day. I really have seen from Webflow that they do still care. Obviously, they still care about the community and they know how many people rely on their platform. They say it all the time. I think we don't really want Webflow continuing to try to pivot and compete with every single new product that comes up and I think that's a good reason to continue to work on their tool. So, the next point is MCP 2.0. So, this was released on July 21st, their new product offering. I have had a little bit of a beta access and I've been thoroughly impressed with how smooth it is and how well it works.

So, it doesn't drop the connection nearly as much as the last version of the MCP did and it also really doesn't require the bridge app. Sometimes my tools will ask to open the bridge app for one or two things that still requires it, but 90 to 95% of the time I'm not using the bridge app at all. It's just connecting in the background. So, that's great. Finally, it doesn't break down all the time as I'm trying to work. I really like how well it connects to components and CMS now. The MCP itself can surface what components you have, what connections are there. You can create properties and you can connect those properties to the CMS right through the interface.

So, through the chat interface with Cursor or Claude or Codex or whatever you might use. And it's distinct because it's not just a code component or an AI component that's locked in a sandbox. That was a problem for me. I really didn't like how the AI components and DevLink components require external code basically. When it goes into Webflow, it's locked inside of a sandbox component and you can't open it up and inspect what's going on. And the reason for that is really it's its own environment. It's a React or an Astro Island or whatever. And the new MCP can create components natively in Webflow that you can hand off perfectly without having to hand off another repo of code.

So, I really like that. I really like that you can bring your own tokens and connect your own tools, whether you're paying for Cursor or Codex or Claude or something else, all of that can connect right into the Webflow MCP. So, you've got all your context, all your skills, everything that you might need to work on Webflow site, you can just bring along with you instead of waiting for or relying on the Webflow AI, whatever it is that lives inside the Webflow system. I haven't had a lot of success with that. I really would much rather prefer to bring my own tools. So, that's possible.

It actually works now. One of the new things as well is you can use your MCP connection to write skills and rules directly onto the website. There's a new section in the Webflow sidebar that contains agent skills and agent rules. And that's great because that's another yet another thing that I don't have to hand off to my clients. I don't have to worry about a code repo or a document file or a readme file. It's I worry about it, you know, I still can use it, but it's built right into Webflow. So, that's great. I can write the instructions and all the design system rules, all of the naming convention rules, any other tools I might use, I can document them right in Webflow.

So, that's a huge deal. I think we need to get on the idea of documenting and being clear with clients what exactly we have done as agencies. What did we do on this site? What tools do we use? So, I think we should see it as a style guide 2.0 and really invest heavily in documenting and being clear with our clients and the clients' agents. I really like that part of the platform. I think it's going to be a big improvement. One of the big things about the Webflow MCP that I think a lot of other tools haven't figured out quite yet is user access and governance.

You know, governance is their corporate word, but in my head it's a great way to just make sure that the right people can edit the right things. And Webflow has already solved this on the user access front because they have different seats and custom roles for enterprises. And the great thing is that your MCP connection respects that role. Whatever role you have or your client has, that's the same role that their MCP and agent connection will have. So, you don't have to worry about somebody who doesn't have publish authorization accidentally connecting their agent, making a bunch of breaking changes, and then publishing the site.

If the person can't make the change, the robot can't make the change either. So, that's great. One of the other things that's cool is the ability to query the audit log with the MCP, and you get a plain text result. You can ask questions. That workflow has been needing a lot of help for a long time, so I'm glad to see the changes and the updates there. And you can also see which change was made by a human versus the change that was made by an agent. Thank you, Webflow, for including that. I think that's a huge deal. Hopefully, one day you don't log in and see your client's MCP connection making a bunch of breaking changes, but at least you'll be able to track it down and explain exactly what happened.

My last point is that clients have different needs. Not every stack is right for every client, and it's pretty clear to me that vibe coding and Vercel and React, that type of stack is also not right for every person. For goodness sakes, we still have people using Drupal in 2026. It's just clear to me that everybody's going to want some specific tool or cleanup or feature that not every platform has. And like I said before, Webflow should not be trying to be every single platform. I think it's got a good identity. And I think people will choose Webflow for the right reasons. They'll choose Framer for other reasons.

They'll choose Sanity for other reasons still. That's important to me that people are able to choose their own stack. It's important that Webflow has the right tools, and I think they're building those tools quite well. I think Webflow is also doing a good job of balancing the concerns of AI and security risks, governance risks with their existing platform. So, we don't just have to vibe code a full user management system for what clients can connect to, how they deploy things. All that stuff you have to manage on your own if you're starting a new site on a new stack. And Webflow has all that built in.

They've had it built in for years, and we don't have to therefore worry about any of that complexity. One of the other things I've noticed with my clients is that non-technical marketing people do not want more complexity. They don't want to learn how GitHub works. They maybe would want to connect their Claude, their desktop version of the app, and get some basic reports or have some basic SEO conversations, but they don't want to be in charge of the marketing site. They want a developer like me to do some of that technical complexity. And so it's nice that I have the tools now that I need with better MCP connections and more advanced and more robust tools that Webflow provides, but some of my clients are hardly even using all the features of Webflow.

I don't want to introduce a bunch of brand new complexity and a bunch of new external code and tools just to make my life easier. My goal is to make their life easier and less complicated. I think that's important. I don't think we should be pushing our clients to a more and more technical platform just because it's maybe easier for us to do the vibe coding nature. So, maybe an AI forward company might want that. And a lot of my clients are not even there on the maturity scale. So, as a recap, I think number one, Webflow hasn't made any big mistakes that make me feel like I shouldn't be building in their environment anymore.

Sometimes they don't do things that I would do or they may prioritize enterprise over agencies and small business, but I get it. They have to make money. So, I don't think they've made any big strategic or structural mistakes. Next, I think the MCP is actually really great now. I'm happy to keep using it. I've been on the beta for a couple weeks and it's been really a great experience. I'm excited to see what people build with it. And number three, I think clients just have different needs. So, don't feel like everybody should be on the same stack. Even within Webflow, I have a lot of different setups and a lot of teams that want to be more or less involved.

So, I still think Webflow is a great platform and I'm excited to see what everybody builds with the new MCP.