— Introducing

AI that lives on your desktop.

Lives where you work. Sees what you see. Acts when you ask.

— The idea

The interface disappears. The thinking stays.

An ambient assistant. Next to your code, your terminal, your tabs — quiet until it has something worth saying.

What makes Lumen different

01

Speaks first, asks second.

Reads what you're working on. Suggests the next step before you can frame the question.

. from any app

02

Lives in your context.

Files, browser tabs, terminal output — Lumen sees what you see, so you don't translate.

current file selection terminal clipboard

03

Learns your shorthand.

The more you use it, the less you have to say. Custom commands, your phrasing, your stack.

rfc opens RFC template
ship it runs CI, opens PR

— In practice

Open it. Ask it. Move on.

Live demo — press K from any app.

you
Where did we drop the rate-limit middleware?
lumen

Found it — server/middleware/rate-limit.ts:42, removed in PR #4821 when you switched to edge runtime.

server/middleware/rate-limit.ts- import { rateLimit } from './rate-limit'
- export const middleware = rateLimit({ limit: 100 })

Want me to re-enable it for /api/* only?

you
Yes, draft the PR.
lumen
  • Created branch fix/rate-limit-api
  • Re-enabled middleware for /api/*
  • Opened PR #4877 → main

Build with someone in the room.

Lumen is free while in beta. Takes 20 seconds to install.

Free during beta · No credit card · macOS & Linux · Open beta closes June 30

Watch the 90s demo