HBR · March 2026
The difference isn’t the model. It’s matching the cognitive lens to the moment.
AI for your work. Meseekna for your thinking.
Four cognitive lenses sit behind great AI use — how you source information, structure tasks, consider alternatives, and reason with numbers. Thinker shows you which lens each prompt of yours drew on, and sharpens the one you choose to grow with a single keystroke, ⌘⇧M.
Your prompts stay on your machine. Always.
Not a prompt optimizer. A coach that watches the prompts you send to AI, scores how you’re thinking, and sharpens the lens you pick — in any text field on your Mac.
Thinker runs quietly on your Mac and reads the prompts you send to AI tools — never the responses, never anything you didn’t type into a prompt field. Each prompt becomes a colored tile in a growing mosaic in your menubar, tinted by the lens it scored highest on. The longer you use Thinker, the richer the picture gets.
Different prompts call for different lenses. Press ⌘⇧M on any focused text field and pick the one this prompt is missing — Thinker writes the version of your prompt framed by that lens. The other three sit back. They get their turn on a different prompt.
Rewrites you accept never count toward your scores — only what you write on your own does.
Friday morning, the weekly recap opens with per-measure cards, the rewrites you accepted, and the mosaic of your week — every tile clickable for the real prompt behind it. The shape of how you thought, on a page.
Both plans run in the menu bar. ⌘⇧M sharpens your prompts.
✓ Cancel anytime. Your prompts and patterns stay on your machine.
Dozens of tools will rewrite your prompts to sound better. Thinker does a different job.
Length, clarity, politeness, formality. The rewrite ships with the model’s preference baked in — usually toward “sounds better.”
None of these tools know which cognitive lens your prompt was missing, because no one has named those lenses or measured them. Every rewrite is some model’s best guess at “better.”
Information, Tasks, Breadth, or Quantitative. You pick the lens the prompt is missing. The rewrite is framed by that lens and nothing else — not all four, just the one the moment calls for.
And Thinker remembers which lenses you’ve been picking across every prompt of your week. The other tools can’t see the pattern. Thinker’s job is to make the pattern visible — so the right lens at the right moment becomes deliberate, not accidental.
It’s the question The Thinker has been holding for over a century — and the question Meseekna’s scientists have been learning to answer since 1965. Not what you think, but how.
Most assessments rely on what people say about themselves — personality inventories, 360s, questionnaires. The result is a measurement of self-perception, not cognition. Meseekna’s scientists built simulations instead: complex decision environments where what mattered wasn’t what participants reported, but what they actually did under pressure. 500+ peer-reviewed publications later, that method is still the basis of everything we build — including the four measures this app applies to your AI prompts.
Thinker runs on macOS — we’ll email you the link so it’s waiting when you’re at your Mac.