Lunem is a private, screen-aware learning companion. It teaches in context, points at what you mean, and remembers your progress — on your device, by default.
Windows 10/11 · Local Ollama by default · Cloud models are opt-in, request by request.
The actual product walkthrough drops soon. Leave your email and it lands in your inbox the day it's up.
Product walkthrough — coming soon
Focused work, screen-aware teaching, and durable memory — built as a single interface instead of three separate tools.
Lunem captures a bounded screenshot, then draws temporary numbered points, rectangles, and arrows directly on your screen to explain what you're looking at. Non-interactive, and it disappears on its own.
Short-term working memory for one goal — task, decisions, blockers, next action — so "continue where I left off" actually works.
A local wake phrase with a visible always-ready indicator. Nothing is recorded or sent anywhere while it's just listening for its name.
Habits, preferences, and decisions are scored, deduplicated, and stored locally — visible and forgettable, never silent.
Ollama by default. Opt into OpenAI, Claude, or Gemini per request — never by default.
An interactive teaching board with model-generated explanation cards and locally persisted ink.
Use "Hey Lunem" or a keyboard shortcut — no microphone button to hunt for.
Recent turns, the active focus thread, and relevant memories are combined. The screen is only captured when the request actually needs it.
Lunem explains, points, or acts — then keeps what's worth keeping, locally, so next time it already knows.
Cloud reasoning is available when you want it — never assumed.
Join the list to hear when the next Lunem build is ready for your machine.