Context
It already knows what you're looking at.
Summon layman and it reads your selection, the window behind it, or the files you picked. It's attached as a chip, and every question in the chat uses it.
Launch sale50% off until October 31
A private AI assistant for Windows and Mac that opens over whatever you're doing. It runs on your computer, so nothing leaves your machine.
Windows 10 & 11 · macOS, Apple silicon & Intel · One-time $29
Say it the way you'd say it. No commands to learn.
01 · What it does
Context
Summon layman and it reads your selection, the window behind it, or the files you picked. It's attached as a chip, and every question in the chat uses it.
Actions
Launcher
Models
Qwen3, gpt-oss, Gemma, Phi and Mistral through Ollama, installed for you during setup. Want a cloud model as well? Bring your own key. layman takes no cut.
02 · Privacy
That's the default, not a setting. The model runs on your computer, and layman only looks at your screen when you press the shortcut.
Choose a cloud model and your messages go to that provider, labelled in the model menu. Your keys stay encrypted on this computer.
03 · Models
layman checks your memory and graphics card, and only lists the models your computer can run.
A rough guide. On your computer, layman measures the real numbers.
04 · And also
;sig in any app and it expands.05 · Roadmap
v1.2. A redesigned launcher, loose wording understood, suggestions with reasons, snippets, calendar, integrations, voice.
Select a paragraph, ask for a rewrite, and the rewrite goes back where the paragraph was.
“Tidy my downloads and zip last month's screenshots.” layman plans the steps, shows you the list, and runs it once you approve.
06 · Under the hood
07 · Pricing
Launch sale50% off until October 31
Cheaper than two months of the AI subscription you forgot to cancel.
Checkout by Lemon Squeezy. Your license key arrives by email.
From the guides
08 · Questions
Yes. Once setup has downloaded a model, layman answers with no internet at all. It needs the internet for web searches, cloud models (if you add one), and a license check every few days. It keeps working offline for up to 30 days between checks.
No. The default model, Qwen3 4B, is 2.3 GB and runs on ordinary laptops. With a good graphics card, or an Apple silicon Mac's shared memory, bigger models get comfortable too. layman checks your hardware and marks each model Runs fast or Runs slower.
layman itself, plus Ollama, the free local AI engine. Setup downloads Ollama's official installer, checks its signature, and installs it for you. Then it downloads the default model. On a Mac it also asks for the Accessibility and Automation permissions it needs to read your selection.
Simple things you asked for, like volume or opening an app, just happen. Power actions (sleep, restart, shut down) ask first. Anything that needs a command is shown to you and waits for Run. Commands never run as administrator.
Yes, with your own API key: OpenAI, Anthropic, Google, xAI, Mistral, DeepSeek, Groq and OpenRouter. You pay the provider directly. layman adds no markup. Signing in with a ChatGPT or Claude subscription isn't possible: neither company allows it for third-party apps.
On the old one, open the dashboard and choose Profile → Deactivate this PC. Then activate on the new one. Old computer lost or dead? Reply to your purchase email and we'll reset the key.
One license covers one computer at a time, and it can be either. Need both at once? That's two licenses.
It's in your purchase email. You can also look it up on Lemon Squeezy's order page with the email you bought with.
Because it explains things in plain words, and you shouldn't need to be an expert to use it. The logo is an “l” and a full stop. That's the whole joke.