Flagship · Early access

Put it in the hopper. We'll tell you when it's done.

Hopper is the layer between you and your agents. It launches your whole setup, watches your agents work in parallel across Cursor profiles, and pings you the second one needs attention — so you stop tab-checking and start switching only when it pays off.

Flagship · Early access
Requires Cursor IDE (free or paid) installed on Windows.
Demo — Morning launch

One command, your whole day boots up.

Open Hopper, hit Launch, and your entire working setup comes alive — four Cursor windows bound to four projects, a Claude Code terminal, your browser preview on the second monitor, your reference tabs. Windows auto-tile into a preset layout of your choosing. No more Monday morning manually dragging windows for ten minutes while the focus drains away.

Forge — Workspace · default-quad1 doneCursor · auth-service✓ doneCursor · web-frontendrunningClaude Code · refactorthinkingCursor · billing-apiidle
Workspace tiling preview
Demo — Agent status

Know when to switch, without looking.

While you're heads-down in one window, Hopper watches the other four. When an agent finishes its task, your taskbar icon pulses orange. When an agent is blocked waiting on you, it flashes. When it fails, it stays red. You never have to tab-check to find out if the agent is done. Your focus stays where it is until the moment switching pays off.

heads-down on the one tile that mattersauth-serviceagent finished its task
Quiet taskbar status
Demo — Multiple Cursor profiles, one launch

Three Cursor profiles. Side by side. One glance to know who's done.

Assign different profiles to different roles — architecture, testing, docs — and launch them together as part of one workspace. Hopper tiles the windows, tracks each one's status, and blinks your taskbar the moment a profile's agent finishes or gets stuck, so you're never silently waiting on a window you forgot about.

Hopper · 3 Cursor profiles running side by sideABCProfile A · architecturerunning — auth refactorProfile B · testing✓ finished — taskbar pulsing orangeProfile C · docsrunning — API reference passLaunch, watch, and get pinged — one window per profile, no silos
Multi-profile launch board
Demo — Built-in tools

Your saved commands, one hotkey away.

A built-in tools page for the shell, PowerShell, Python, and Node one-liners you run all the time — plus quick links. Save them per-project or globally, run them with one click, and see stdout/stderr right there. Agents can register tools here too, so a script an agent sets up shows up ready to run for you next time.

tools · saved commands, one click awaytest:authrun auth test suitenpmdeploy.ps1push build to stagingpowershelllint-fixeslint --fix on src/shellopen-dashboardjump to staging dashboardurlSave once, run anywhere — you or a registered agent tool
Built-in tools console
Demo — Ambient startup music

Your workspace launches. So does the soundtrack.

Point Hopper at a local audio file, and it plays automatically the moment your workspace boots up — one less click between you and heads-down work. No streaming account, no window to manage. Just a startup ritual that puts you in the zone before the first prompt.

main.ts — CursorNOW PLAYINGstartup-loop.mp3local file · your libraryplays automatically on launchReference
Workspace + ambient startup track
Demo — Your own dashboard

See how you actually work.

Every launch, every agent completion, every context switch is tracked to a local SQLite database on your machine. Hopper gives you a dashboard showing how long sessions typically run, how often agents finish vs. get stuck, which times of day you're most productive. No telemetry leaves your computer. It's for you, not for analytics vendors.

Hopper · Analytics — last 30 days100% localSessions per day · agents finished greenSessions186+24%Agents shipped94+38%Avg session47m−12%Top toolCursor64%
Local-only analytics