Ask AI about anything on your screen. One hotkey captures your display and gets an instant answer from Claude or GPT — right on top of your current app. No screenshots saved, no copy-pasting, no tab-switching. Free plan: 10 queries/month, no key needed. Pro: 200 queries + BYOK unlimited. Screenshots stay in RAM only — never stored. Supports Claude Haiku, Sonnet, Opus & GPT-4o. macOS 12+.
Hey Product Hunt! I'm Krunal, the maker of ScreenGeany AI.
The problem that bugged me:
Every day I'd find myself doing the same thing — see something on screen, screenshot it, open ChatGPT, upload it, wait, then switch back. For a 10-second question, it took 45 seconds of context-switching.
So I built ScreenGeany AI.
One hotkey captures your screen and gets an instant AI answer — right on top of whatever app you're using. No screenshots saved. No copy-paste. No tab-switching.
How it works:
- Press the hotkey (customizable)
- It captures what's on your screen
- Ask your question
- Get an answer from Claude or GPT-4o — without ever leaving your current app
Some real ways I use it daily:
- Stuck on a form? Ask what a field means
- Reading docs? Ask for a summary
- Debugging? Ask about the error on screen
- Reviewing a contract? Ask about a clause
What's included:
- Free plan: 10 queries/month, no API key needed
- Pro: 200 queries/month + BYOK for unlimited
- Your screenshots are never stored (RAM only, deleted immediately)
I'd love to hear how you'd use it — drop a comment! And if you have feedback or feature requests, I'm right here.
Report
How does the app ensure screenshots are fully cleared from RAM after each query, and is there any independent verification of this privacy claim?
Screenshots are captured as NSImage objects held only in memory (RAM). Once the AI responds, the
reference is released and the memory is deallocated — no file is ever written to disk.
The image is sent directly from your machine to the AI provider (Anthropic/OpenAI) over HTTPS. It
never touches our servers.
Really appreciate you asking — thanks!
Report
the 45-second context switch problem is real and annoying. the "stuck on a form field" use case is the one i'd lead with in marketing - it's the most universally relatable. curious whether you're seeing people use it more for quick lookups or longer analytical questions like contract review. congrats on the launch krunal.
Ivo Gospodinov Thanks Ivo! The "stuck on a form field" moment is exactly what made me build this — kept running into it myself. Great call on leading with that.
To your question — it's day one so early to say, but right now it's mostly quick lookups ("what does this mean?", "draft a reply"). Longer stuff like contract review works too since you can ask follow-ups.
This is a nice, simple idea. The “ask about anything on your screen” use case comes up a lot, and having it triggered by a hotkey directly on top of the current app feels pretty natural.
Keeping screenshots only in RAM is also a good call. A lot of people hesitate with screen tools because of privacy concerns. The real test will probably be how fast and accurate the responses feel in day-to-day use.
@konstantinalikhanov Thanks Konstantin! Privacy was non-negotiable — screenshots stay in RAM, encrypted over TLS 1.3, deleted the second AI responds. Nothing stored anywhere, ever.
On speed — responses start streaming in about 2 seconds, and you can switch between Claude and GPT-4o depending on what works best. Give it a try and let me know how it feels!
Report
useful initiative, would be nice to be able to use it on the web version without having to download the app but I understand that there are limitations with that.
ScreenGeany AI
How does the app ensure screenshots are fully cleared from RAM after each query, and is there any independent verification of this privacy claim?
ScreenGeany AI
@mordrag Great question, Denis!
Screenshots are captured as NSImage objects held only in memory (RAM). Once the AI responds, the
reference is released and the memory is deallocated — no file is ever written to disk.
The image is sent directly from your machine to the AI provider (Anthropic/OpenAI) over HTTPS. It
never touches our servers.
Really appreciate you asking — thanks!
the 45-second context switch problem is real and annoying. the "stuck on a form field" use case is the one i'd lead with in marketing - it's the most universally relatable. curious whether you're seeing people use it more for quick lookups or longer analytical questions like contract review. congrats on the launch krunal.
ScreenGeany AI
Ivo Gospodinov Thanks Ivo! The "stuck on a form field" moment is exactly what made me build this — kept running into it myself. Great call on leading with that.
To your question — it's day one so early to say, but right now it's mostly quick lookups ("what does this mean?", "draft a reply"). Longer stuff like contract review works too since you can ask follow-ups.
Would love to hear what you'd use it for!
Banyan AI Lite
This is a nice, simple idea. The “ask about anything on your screen” use case comes up a lot, and having it triggered by a hotkey directly on top of the current app feels pretty natural.
Keeping screenshots only in RAM is also a good call. A lot of people hesitate with screen tools because of privacy concerns. The real test will probably be how fast and accurate the responses feel in day-to-day use.
ScreenGeany AI
@konstantinalikhanov Thanks Konstantin! Privacy was non-negotiable — screenshots stay in RAM, encrypted over TLS 1.3, deleted the second AI responds. Nothing stored anywhere, ever.
On speed — responses start streaming in about 2 seconds, and you can switch between Claude and GPT-4o depending on what works best. Give it a try and let me know how it feels!
useful initiative, would be nice to be able to use it on the web version without having to download the app but I understand that there are limitations with that.
ScreenGeany AI
@viktorgems Thanks Victor! You're right — a web version would be convenient, but the core experience really
depends on being a native app. Here's why:
The magic is the global hotkey that works across ANY app — your browser, IDE, PDF reader, email
client — without switching context. A web app can only see its own tab, not your whole screen.
That said, we're exploring a Chrome extension that could cover browser-based use cases (forms,
emails, articles). Would that be useful for your workflow?
Appreciate the kind words and feedback!