Glaze by Raycast

I don't understand how these type of projects are still tried and get any traction... anyone who has tried them will 100% know it won't go beyond a happy path demo. If they want to seriously use/publish the app beyond playing around, it'll require weeks of iteration via AI, which will cost you an arm and a leg in tokens.

It's certainly a nice promotional website.

My first thought was, "So, Replit and ilk?", seems they expected that comparison:

> How is Glaze different from Lovable, Replit, or v0?

> Those tools build for the browser. Glaze builds for your desktop. That means your apps can access your file system, your camera, keyboard shortcuts, menu bar integration, and background processes. Things a web app can’t do. It’s a different category entirely.

Pretty sure modern web apps can do all of those (sans menu bar). (If anything they do background processes better since you can send a very long task off to a server and shut off your computer, come back later and pick up where you left off.)

Also, as others mentioned, this just seems like Claude Code with extra steps, unless they managed to nail some sort of design standard enforcement they feel is better than what most people can get out of it.

The quick publishing is kind of nice, but it immediately made me think it would be more interesting to have a way to quickly remix other people's creations, similar to the Figma Community tab: you can take someone else's work, break it apart to see how it works, then tweak it how you want it.

Ironically, there's another project named Glaze, that aims to "protect artists from generative AI" (https://glaze.cs.uchicago.edu/)

I cannot bring myself to trust unreviewed software enough to install it on my own machine with arbitrary permissions. I understand the push for AI-generated websites, because the code running in my browser's sandbox is gonna have very limited permissions to do anything evil, but desktop apps are a completely different story

Might sound like a rube here, but: is agentic development really this good at novel UIs? The video shows a sort of cassette tape music player, and a fancy looking audio visualizer/equalizer thing. I'm well aware agents are very good at boilerplate UIs, but I wouldn't expect them to be able to one-shot novel, dynamic UI elements like this. I've had Claude attempt some SVG animations and the results were very crude. That was a year or so ago though. Are there established ways of letting agents iterate on UIs, i.e. having them visually verify the visual design and interactions?

Not to be a curmudgeon, but why are they spending time on this? As an enthusiastic Raycast user, I would prefer to see them focus on making Raycast better, not finding new ways to jump on the AI bandwagon.

Claude Code is pretty good at Swift + Swift UI. I created and have been iterating on a menubar app for myself that I plan to share with a small team. I'd prefer to do this native than go through a 3rd party solution.

I do have prior experience developing for iOS but that was pre-swift.

Someone please exfiltrate their prompts/skills so i can use these on Codex. I've have relative success building my own apps for mac using Codex but they're uglier than sin and dont seem to understand well how permissions work.

Having antigravity is useful because Gemini 3.1 is pretty good at generating UI sugar. Claude 4.6 Opus provides nothing to write home about. Their shadcn looksmaxxing hasn't generalized to writing good desktop UIs.

Raycast's only edge here seems to be the fact that they are obviously very good at Mac app development and probably have impeccable skills/documentation for building them.

Taking a step back, it's pretty clear that Raycast is angling for an Apple acquisition here with this play. If I'm Apple, the reason to buy a product/team like this is a no-brainer.

But can it manage external libraries or use only the existing sdk? I had a non tech friend run into an issue recently where she wanted to automate a pdf action. Eventually I realized she needed to run homebrew and install a library. Curious if this actually manages that kind of process.

My metric for this kind of stuff is: Did Glaze build the Glaze app?

Just what absolutely no one needed: another locked down and non web platform with horrific security that tries to digitally enslave people just the tiniest level above what they can accept now. I don’t see any future where raycast can survive and i would say its a good thing.

Looking forward to trying this out and see how this differs from more manual approaches. One thing that stands out is an included store for public/private distribution — that’s super convenient given the cumbersome (and maybe, horrific) process that is the app store submission.

Getting closer to ClaudeVM https://jperla.com/blog/claude-electron-not-claudevm

So it looks like they’re creating their own App Store within the app? At least it’s kept separate from official apps. But also how is that not a security nightmare Apple won’t allow?

This looks super fun, actually.

I wonder what it is actually building. Tauri apps, maybe?

It is interesting how so many different companies end up converging to some sort of AI coding.

Raycast -> Glaze AirTable -> Lovable Competitor Retool -> Lovable Competitor

Even those early in the journey are converging towards coding.

Have had good results on MacOS just using codex (or your cli of choice).

Have it create a swift app, unless extended permissions are needed it can compile withouy going into xcode.

Few simple util apps, disk cleaner, clipboard manager. Worked pretty well.

Had better results than using xcode's built in ai extension.

There seem to be more AI app building platforms than actual apps being built these days.

No mention of security.. remarkable

I can't imagine trusting these apps with access to my camera, file system or any other sensitive permissions.

Impressive feat. Definitely not for me though, and for sure I won't be there to debug one of these when my parents will call me because it broke their computers.

This is just a landing page. There's not even any decent product specs. Nothing technical. How does this make front page of hacker news?

It will be awesome if these were native apps instead of JavaScript apps. It's not mentioned anywhere explicitly that these are native.

How many apps do you really need that are not already done - perhaps even better?

So, another wrapper around claude 4.6 for +xx% higher price? Using just claude code, one can do what glazeapp seems to aim for, no? "Beautiful by default" seems to be a system prompt akin to:

  Design Philosophy Create apps that feel premium, polished, and worthy of being featured on Dribbble's most popular shots. Every pixel matters. White space is your friend. Less is more, but what remains must be perfect.

  Visual Design Principles

  Color & Theming

  • Use sophisticated, limited color palettes (2-3 primary colors maximum) ...

so many unknowns...

1. macOS and Windows require installation of Xcode and Visual Studio respectively, and if in Apple's case you kinda can install these tools headlessly and choose to install only the "build tools" package, Microsoft's creature is gonna daze and confuse you with a crap-ton of checkboxes and no easy "just install whatever is minimally needed to compile my code" button, and I don't recall if there is way to install build tools on Windows through terminal.

2. what is going to be distributed? source code itself or actual binaries? and what will the security model of Glaze store be? same as extensions, "everything is open-source and undergoes Raycast's and community review"?

3. Glaze is going to come to Windows and Linux, if we trust the Q&A section at the end. what will Glaze build upon? separate frameworks and languages for each platform or something multi-platform [1] like Tauri or Kotlin Multiplatform? or are you going to copy the Raycast extension model - just run Node, expose some platform integration, and parse React render trees through "Glaze Runtime"? I've been working on a bug in Vicinae [2][3], and I've seen this model in action. it's very hard to make it perform well, but all it takes to achieve native look and feel is to just map React render trees to whatever system component OS offers. (in Vicinae's case, it's Qt. bet that it's done with SwiftUI on macOS and WinUI 3 on Windows.)

[1]: there is a difference between "cross-platform" and "multi-platform". "cross-platform" means "I behave equally across platforms and have no awareness of native look and feel" (e.g. Electron, Unity, Flutter), while "multi-platform" means "I can adapt across platforms to the degree you need" (e.g. C/C++, Rust, KMP)

[2]: https://github.com/vicinaehq/vicinae

[3]: https://github.com/vicinaehq/vicinae/pull/1158

what about barebone/starter desktop app that can be modified itself by prompts?

that's would be Electron app, but without unneeded bloat

Vibe-coding desktop apps is a much, much better solution for the vast majority of one-off tools most users want to build.

No thanks.

As an interesting counter-proposal to wasting time with this... look for older less popular/downloaded/featureful apps written by people for their own education, edification and enjoyment.

They may not work the way you wish they would, but you can learn a lot from them, be inspired by them, and leave feedback.

That's how you actually encourage more people to get started and continue making their own tools.

they did it again, glad I am on Mac, congrats raycast

I love Raycast. I would probably be called a "power user" - I use it all day long and have a fairly sophisticated and customized configuration and set of workflows. Raycast is actually one of the primary things keeping me on MacOS these days (please release a Linux version!).

I am worried this is the start of them trying to diversify their product offering because revenue has stalled in the core Raycast product and VC demands more returns. I don't want to be jaded, but history teaches me to be. Here's hoping that Raycast itself is still a focus for the company.

Is this just shitting out electron crap?

"Insecure apps, reimagined by you"

I thought this must be a joke at first. "Glaze" is in pretty heavy use as recent slang for "when someone excessively praises another person in a way that feels over-the-top." https://creativesimiles.com/glaze-meaning-slang/

ie the annoying way that LLMs interact with users

[dead]

[dead]