The feature I kept forgetting to talk about
Every time I show someone LacPointer, I demo Option+Space, voice control, maybe the integrations. Then I hit Caps Lock and their face changes. That's Wand. And somehow it never gets enough attention.
Wand is simple: press Caps Lock on your keyboard and hover over anything visible on your screen. An error message in a terminal. A pricing table in a browser. A dense paragraph in a PDF. A weird icon in a UI you've never used before. LacPointer reads whatever is under your cursor and sends it straight to the AI for an explanation.
No selecting. No copying. No opening a tab and pasting. You just point.
What "under the cursor" actually means
When Caps Lock is active, LacPointer uses accessibility APIs to grab the text or element context from whatever is sitting beneath your cursor at that moment. On macOS this goes through the Accessibility framework. On Windows it works through UI Automation. It's not a screenshot OCR hack — it's pulling real text from real UI elements, so the quality is much better than you'd expect.
That means it works in places you wouldn't assume: inside Electron apps, across most native windows, inside browser tabs. It even picks up things like button labels, tooltip text, and error codes that you'd normally have to squint at before remembering to Google.
Copy Mode: the passive version of Wand
If hovering feels like too much friction (it won't after a day), there's Copy Mode. Enable it and LacPointer watches your clipboard. The moment you copy something — anything — it automatically explains it. No Caps Lock needed, no hover. Just Cmd+C or Ctrl+C like always, and the explanation appears.
I use Copy Mode when I'm reading documentation or digging through logs. I'll copy a stack trace line, and before I've even thought about what it means, LacPointer has already broken it down. It's one of those features that sounds minor until you've used it for a week and then can't go back.
Where I actually use Wand every day
A few real situations from this week alone:
- Error messages in the terminal. Instead of copying the error, opening a browser, pasting into a search bar, and skimming Stack Overflow — I just hover over the error and press Caps Lock. The explanation is right there in LacPointer's floating bar.
- Unfamiliar UI in a tool I've just installed. Hovered over an unfamiliar toggle in a settings panel I'd never opened before. Wand told me exactly what it controlled. Saved me from reading a manual I wouldn't have read anyway.
- Code I didn't write. Hovering over a function signature in someone else's codebase and getting a plain-English breakdown is genuinely faster than jumping to definition and reading the implementation.
- Legal or policy text. I know, I know. But when you're staring at a terms of service clause or a privacy policy section and you actually want to understand what it's saying, hovering with Wand beats re-reading it three times.
Sharing results and pushing to Notion
Once Wand gives you an explanation, you're not stuck with it sitting in the LacPointer bar. You can share the result with teammates directly, or push it to Notion instantly. That second option has become part of how I document things I learn on the fly. See something confusing, Wand explains it, push it to a Notion page titled "things I learned this week." Takes about four seconds total.
The Notion push goes through the same integration that powers the rest of LacPointer's Notion features. If you've already connected your Notion workspace to LacPointer, there's nothing extra to set up for Wand results.
A few things worth knowing
Caps Lock as the trigger is clever once you think about it. It's a key most people never use intentionally, it's reachable without looking, and it doesn't conflict with basically any app shortcut. When Wand mode is active, your actual Caps Lock (the uppercase toggle) is temporarily suspended, but it comes back the moment you're done.
The explanation quality depends on which model you've configured in LacPointer. If you're on a strong model like Claude or GPT-4o, the explanations are detailed and context-aware. If you're on a lighter model, they're still useful but shorter. You can have up to 3 API keys on the free plan and up to 10 on Pro, so it's straightforward to plug in whatever model you prefer.
Also worth noting: Wand works regardless of which Persona you have active. If you've got the Coding persona running, the explanation will come back in that voice — direct, technical, no hand-holding. If you've switched to the default persona, it'll be a bit warmer. A small thing, but it means Wand fits into whatever context you're working in.
Why this matters more than it looks
The thing about Wand is that it collapses the gap between seeing something and understanding it. That gap used to involve at least three or four steps — copy, switch, paste, read, come back. Wand makes it one step. And when that gap disappears, you stop losing your train of thought every time you hit something unfamiliar.
I've noticed I stay in flow longer on days I'm using Wand actively. That sounds like something you'd put in a marketing brochure, but it's just the practical result of fewer context switches.
Try it right now
If you've got LacPointer open, press Caps Lock and move your cursor over something on screen you haven't thought about in a while. An icon in your menu bar. The text in an app you use every day but never really read. A number in a dashboard you've been ignoring.
If you want the passive version, go into LacPointer settings and flip on Copy Mode. Then just work normally for the next hour. You'll notice it doing its thing every time you copy something.
And if you haven't downloaded LacPointer yet, it's at lacai.io/download — free tier gets you 4,000 tokens every 15 hours, which is plenty to get a feel for Wand without spending anything.