Should I vibe code
Capture screenshots or screen video, annotate them, and copy or save locally
Capture, annotate, copy. Every piece of this is a platform API you already have.
?
Their verdict, the One-time license price and the build-time estimate come from their entry, MIT-licensed. Checked 2026-08-03.
?
Our verdict, the regret score and everything below it. Editorial and unsponsored — nobody can pay to be moved.
The honest answer
why the verdict is what it is
Screenshot tooling sits right in the sweet spot: visible payoff, no persistence, no other humans involved.
What actually breaks
not "if". the specific failures.
- Screen recording permission, which macOS revokes on update and which fails by producing a black image rather than an error
- Capturing at the wrong scale factor, so a Retina screenshot is saved at half the resolution it was taken at
- Multi-display capture, where a window spanning two screens with different scale factors is captured at one of them
- Annotations baked into the pixels, which cannot be moved or removed once saved
- Whatever else was on screen, since a full-screen capture takes the notification that arrived a second ago with it
The realistic failure is a black rectangle. macOS updates, screen recording permission is silently reset, and your tool keeps working — the hotkey fires, a file appears, the notification says it saved. The file is entirely black, because the API returns empty frames rather than an error when permission is missing. You discover this when you go to send the screenshot you took twenty minutes ago and have already dismissed the thing you were capturing.
Is that you?
the verdict is a default, not a law
- It captures at native scale and you have verified the output resolution
- Annotations are kept as a separate layer until export
- It is local, and captured images stay on your machine
- You do not detect and report missing screen recording permission
- Captures are saved at logical rather than native resolution
- Annotations are irreversibly flattened with no editable copy
- It uploads captures anywhere by default
If you build it anyway
the checklist, then the prompt that enforces it
- Check screen recording permission explicitly on every capture and fail loudly. The API returning black frames instead of an error is the defining trap of this category.
- Capture at native backing scale and verify the pixel dimensions match the display, not the logical points.
- Keep annotations as a separate editable layer, and only flatten on export — a baked arrow in the wrong place means recapturing.
- Handle per-display scale factors when a window spans screens, and capture the window rather than a screen region where possible.
- Suppress notifications during capture, or at least warn, since a full-screen shot takes whatever arrived with it.
- Strip metadata on export and keep everything local by default — not uploading is the reason to build this.
Before you build a screenshot tool, apply these and push back if I ask you to break them. 1. Check screen recording permission before every capture and fail with a visible message. Tell me the platform API returns black or empty frames rather than an error when permission is missing, and that macOS resets this on updates — so the failure is a saved file that looks like success. 2. Capture at the display's native backing scale. Verify the output pixel dimensions against the physical resolution rather than logical points, and tell me when they disagree. 3. Handle multiple displays with different scale factors, including a window that spans two. Prefer capturing a specific window over a screen region. 4. Keep annotations as a separate editable layer in the working document and flatten only on export, so a misplaced arrow does not mean recapturing. 5. Suppress or warn about notifications during capture — a full-screen shot takes whatever appeared with it. 6. Strip metadata on export. 7. Keep everything local by default. Do not add upload or cloud sharing unless I ask, and if I do, make it explicit per capture rather than automatic. 8. Handle the clipboard properly: copying should place a real image, and saving should use a predictable naming scheme I can configure. 9. Out of scope unless I ask: scrolling capture, OCR, video editing, cloud history.
That one keeps you out of trouble. For the prompt that actually builds it, canivibecodeit.com has one.
their build prompt ↗Or don’t build it
the boring option, and the way back out
CleanShot X is a one-off purchase and very polished. Build your own because every piece is a platform API you already have, as the hot take says — the only thing that will bite you is permission failing silently, and checking for it is a few lines.
Captures are just image files in a folder, which is the whole appeal. Keep the annotation working documents alongside them if you build layered editing, and the tool is disposable without losing anything.
Open-source screen recording and sharing product with modern desktop architecture.
Questions
Why does missing permission produce a black image instead of an error?
Because the capture API still succeeds — it returns frames, they are simply empty. From the code's perspective everything worked, so a naive implementation writes the file and reports success. Checking the permission state explicitly before capturing is the only way to turn that into an honest failure.
Why keep annotations as a layer rather than drawing on the image?
Because annotation is iterative in practice — the arrow is slightly off, the highlight covers the wrong line. Flattened immediately, fixing any of that means taking the screenshot again, and the thing you captured has usually moved on. A layered working document costs little and removes the most common frustration with these tools.
Every week, someone ships something they shouldn’t have.
New verdicts, the worst thing that landed in the trap, and the occasional incident report. No other email, ever.
last reviewed 2026-08-03 · verdict is editorial and unsponsored · shared entry data from canivibecodeit under MIT · not legal advice