Leaving Webflow for Astro: Q&A with Alex Glevsky
Alex Glevsky is #1 in our main ranking of Webflow freelancers with 9.80 out of 10. He is a Webflow Certified Partner who has worked on Webflow since 2013 and now moves Webflow sites to Astro, set up so the client's team edits them through Claude Code. Below are his answers to the questions buyers ask most about leaving Webflow.
Why are teams looking for an alternative to Webflow?
Webflow is a genuinely great platform, and I have worked with it for years. What I keep seeing is that nobody on the client's team uses the Designer. There is a learning curve, and people would rather not spend their time on a tool.
Components help up to a point, but on a large project you still end up coming back to a Webflow developer, because every new page differs from the last one. Meanwhile the same people put a page together through Claude or ChatGPT with no trouble at all.
Can't you just use AI inside Webflow?
Through the Webflow MCP, yes, and with some persistence you get results. But it is a layer on top of the existing interface, a long way from working with Claude directly. Anyone who says it changes the process has probably not tried to build anything real with it.
What is Astro, in plain terms?
Any website ends up as a set of HTML files on a server, Webflow included. Astro is one way of organising those files: it lets you reuse what is already built, keep content structured and hold the project in order. After a while you forget it is there.
Who should move, and who should stay on Webflow?
A new site of up to thirty to fifty pages, with a blog updated every couple of weeks, does not need Webflow. Astro is a full alternative for that.
A large Webflow site with dozens of collections and hundreds of pages is different. The move is slow and things can get lost. There I would rather find what is not working and fix it inside the current setup than replace the platform for the sake of it.
Will search rankings drop after the move?
I cannot guarantee that, and nobody honestly can. But every URL, title and line of copy stays where it is, so Google sees the same site on a new stack, served faster. That is as close to zero impact as it gets.
How long does it take, and what does it cost?
About three weeks for a normal business site. Most of that goes into the first pages and the design system; the rest of the pages come together quickly after that. It is $2,500 for a small site and from $5,000 for a normal marketing site.
Where does the content live without the Webflow CMS?
In structured files inside the project, which Astro reads and the agent edits directly. That is enough for most sites. If there is a lot of content, or writers who will never open a terminal, a CMS like Sanity sits on top of the same files, and that is worth deciding at the start.
What does editing the site look like afterwards?
You update the site from Claude or ChatGPT and publish from there, with the source in GitHub and hosting wherever you like. Cloudflare works very well for that. You still need to check that what comes out fits the rest of the site, and the skills set up inside the project catch a good part of that.
So where is the human still needed?
At the start. Current models are still some way from working without a person in the loop, and the architecture decisions at the beginning are not theirs to make. Those decisions decide how fast the project is to work on later. Getting that part right is my job; once the foundation is sound, new pages come together quickly.
In short
- Good fit: a new or existing site up to about 50 pages, with a blog updated every couple of weeks.
- Be careful with: large Webflow builds with dozens of CMS collections and hundreds of pages.
- Price and timeline: $2,500 for a small site, from $5,000 for a marketing site, about three weeks.
- After the move: every URL kept, source in GitHub, edits through Claude Code.
Service details: Webflow to Astro migration on glevsky.com. His full profile, scores and rates are in the main ranking.