We reached the limit of browser-based design tools. This is not a guess about the market. It is a wall we hit ourselves, building Splitleaf.
The browser has a ceiling
A browser tab is not a real open workspace. It is a window with defined borders.
A large site fills a tab's memory fast. The tab slows down.
The browser also blocks a web page from your files, on purpose. This keeps you safe from bad sites. It also blocks a good design tool from real, deep file access.
An AI agent that edits many files at once needs a stable, direct line to those files. A browser tab cannot give this. Every workaround adds a new point of failure.
Browsers change their own rules often. A tool built for one browser version can break after the next update. You did not change your workflow. The ground moved under it.
None of this is a small tax. Each limit stacks on the last one. A tool that must behave the same in every browser and every update spends its energy on survival, not on your site.
One outage should not stop your whole business
Cloud design tools add a second problem: they need a live connection to work at all. No connection, no work.
This is not only a problem for the person who relies on the tool for a living. It changes what happens when something breaks.
If the design tool goes down, you cannot publish. If the hosted site goes down, visitors see nothing. Many companies split these two jobs across two vendors, to lower this risk.
But one level up the chain, the split stops helping. In late 2025, Cloudflare, a company some say quietly carries a fifth of the web, went down twice within three weeks. On December 5, the outage took Canva and Webflow offline, for a reason connected to none of them.
You cannot see that layer. You cannot choose around it after the fact. You can only choose, before you build, how much of your work depends on a chain you do not control.
Fast setup can mean a slow trap
A hosted tool is fast to start. Sign up, and you are building in a minute.
That speed has a cost. Your site now lives on the host the tool chose for you, not the host you would have chosen. Moving later means a real migration, not a setting change.
Many teams accept this trade on day one and regret it in year three. We do not think a tool should force that trade on you.
You should not rent your own work
Here is the plain version: with most of these tools, you rent your website. Stop paying the monthly fee, and your site, your content, and your edit history go with it.
We do not think a business should ever face that choice over its own website. A tool can end its support. A tool should never end your ownership.
We sell the App once, not every month
This is why Splitleaf is an app you install, not a page you visit. You buy the App once. You keep it, and you keep using it.
The App changes the economics, not only the mood. A browser-based tool must run a server for every user, every minute, forever. Someone pays for that server, in the price, forever. The App runs on the machine you already own. Our cost does not grow with what you build. You keep the margin a subscription would have taken.
Today, the App does your design and content work on your machine. Publishing and sign-in still need the internet, because a live website needs a live host. Over time, we plan to move more work local, including the AI agent itself. On-device AI is still young. We are building toward that. We are not claiming it today.
Your files. Your computer. Your GitHub.
A Splitleaf site is a folder on your computer, with real HTML and CSS inside it, and real history in a GitHub repository you own. Not a database row inside our servers. Not a format only we can open.
You can move that folder. You can open it in another tool. You can hand it to another person. You can do this whenever you want, for any reason, without asking us first. That is what ownership means.
Simple enough for anyone, with help nearby
We built the setup steps for someone with no technical background at all.
If you get stuck anyway, you do not need to wait for us. Any AI agent you already use on the web can read our guidance. It can walk you through the step. We wrote that guidance for agents to read, not only for people.
The browser is still a fine place to ask a question. It is just not the place we build your website.
What we believe
We do not think this is only a technical choice. It is a stance on who should hold the risk.
A rented tool asks you to trust its uptime, its pricing, and its plans for your market, forever. A local tool, sold once, asks you to trust it for one purchase. After that, the files are yours, the history is yours, and the risk is yours to manage, not ours to hold over you.
We built Splitleaf this way because we would want to use it this way ourselves.
Bernt Roalkvam is the founder of Splitleaf, a local, one-time-purchase app for designing, managing, and publishing websites.