Last Wednesday, I was tweaking a client’s landing page—and juggling software again
At 10 p.m., the client said the button color needed to be softer. I was staring at a static mockup, but I still had to find the page, copy the styles, and wait for it to run. I’ve also gotten stuck in this revision loop: I’d spend two hours changing something, only to hear, “Actually, the first version was better.” If your business comes from building websites or customizing pages, that communication overhead can eat straight into your profit.
What is Onlook? It puts the canvas on top of your real project
In short, Onlook is an open-source desktop app. You can drag elements around on the canvas or describe an idea with AI. The changes go into your existing project’s code instead of producing an image you can’t ship. You can inspect or discard each change individually. It feels like rearranging products on the shelves of a real store, rather than drawing a mockup and asking someone else to rebuild it.
When founder Daniel launched the project in San Francisco, more than 25,000 designers and developers had already signed up, with over 100 community members contributing. It isn’t a “draw once, make money automatically” tool. It’s more useful for people who frequently deliver web projects.
What I’d prepare for one test run today
The software itself is free to use, while charges for its optional AI services depend on what appears in your account. I’d start with one 30-minute test before committing to a long-term plan. Expect to spend about 1–3 hours, depending on whether the project already runs. The technical barrier is basic familiarity with the structure of web files. If you haven’t touched code at all, I’d ask someone technical to walk through the first pass with me.
The first step isn’t installing anything. I’d open the project’s website, download the desktop app, and create a test project. Then I’d choose the simplest landing page, click a piece of text or a button on the canvas, change one color, and inspect the generated changes. This tool isn’t essential for everyone, and there’s no urgency to try it right now.
How I’d introduce it based on where I’m at
If I’m just starting and still taking template-based jobs, I’d first run through it on one of my practice pages. I wouldn’t add it to the official delivery process until I knew it, because learning the tool could otherwise slow a project down.
If I already have one or two stable clients, I’d test it on a small page that needs repeated revisions. I’d compare each visual adjustment with the corresponding code change and see whether it removes a round of rework.
If I’m scaling up and need design and development to move more smoothly, I’d then assess whether multi-person collaboration, version management, and AI usage costs make sense for the team.
I could also start with a screenshot of a page I revise often and work out whether Onlook is worth adding to the workflow.