Author
Recently just before a client presentation, I realised I didn’t actually have an effective way to review a set of full-page website screenshots on my Mac.
Preview, the Mac default image viewing app, works well enough for small images but long web page jpegs quickly become awkward to interact with. You’re forced into either a “fit to screen” view where everything is unreadably small, or a 100% pixel view where you spend your time panning around oversized content. What I really needed was something simple: screenshots displayed at 100% width of a browser, scrollable with the keyboard or mouse, and the ability to move left and right between images/pages.
On a PC, this would have been an easy decision – I’d have opened IrfanView without a second thought and changed the scaling settings. On macOS, I realised I didn’t have a go-to equivalent with the same settings. With an hour to go before the call, I briefly considered the usual process of trying half a dozen apps and abandoning most of them. The alternative was to awkwardly scroll around a set of Figma frames, zooming in and out.
Instead, I made the totally obvious, sensible decision to quickly prompt-code (a term I prefer to “vibe coding”) a tool using Figma Make.
What is Figma Make?
Figma Make is an AI-powered coding tool (based on Claude Code, I believe) within Figma that turns text prompts and designs into working, interactive prototypes and code. Instead of manually building every component, you can describe what you want like a layout, interaction, or UI behaviour and Figma Make helps generate it directly from your design context. It is primarily fuelled by React, Typescript and Tailwind CSS, but it can be convinced to use other technologies if you prompt hard enough.
Panic!
After some initial mild panic, an hour and about 10 prompts later I had a drag-and-drop interface where I could add all my screenshots in one batch, reorder or remove them as needed, and navigate using arrow keys and scrolling. I added the ability to press “n” to drop notes wherever the cursor was positioned, annotate individual screenshots, move notes around, and have everything saved.
Brilliant, but limited
Figma Make has plenty of flaws, which I won’t get into here. But for rapidly creating a tool that does exactly what you need – and nothing more – it turned out to be surprisingly effective. Sharing screenshots via a temporary Figma URL was perfectly adequate for this use case, though it would obviously be more work if you needed to extract the React files and self-host them. For highly confidential work, a Figma URL may also not be ideal, and you do need to be careful not to accidentally enable “Share to Community” when publishing.
It’s telling that creating a disposable, simple task-specific tool with AI can now be faster than searching the App Store or the web for the “right” solution.
It could become costly
That said, Figma Make does require a full paid designer seat, and I’d be very cautious about using it as a broader prototyping tool. Currently the number of credits are not capped for prompting, but after March 2026 a limit of 3000/month will be imposed. According to the current credit estimator tool in Figma Make, some of my test prompts were taking over 250 credits at a time. So, over-reliance on such tools could become expensive to experiment and work with. It is also extremely easy to create something that looks convincing but could be costly to turn into production-ready code.
As has been quoted to death:
With great power comes great responsibility.