A macropad for Claude Code

One tap.
Minutes of work.

Nine keys wired into Claude Code. Press one and a full instruction is typed for you — review a diff, ship a PR, audit security, plan a feature — work that would take minutes of typing, fired in a single click.

51 ready shortcuts0 setuppick any 9
NOVEMA

How it works

From a key press to a finished task.
01 / BASICS
CLAUDE CODE

The agent in your terminal

Claude Code is Anthropic's command-line coding agent. It reads your project, runs commands, writes and fixes code — you direct it in plain language.

THE KEYPAD

A dedicated remote for it

Each key tells a small desktop app to type a command — or a whole instruction — into the focused Claude Code session, then optionally press Enter for you.

A SHORTCUT

A command, or a full macro

Some keys send a built-in slash command like /code-review. Others send a paragraph-long brief that kicks off a real multi-step run — no custom files, nothing to configure.

press → types the instruction → ↵ → Claude works

The shortcut library

Each card shows exactly what the key types. ↵ fires immediately; “leaves cursor” waits so you can add a target.
02 / 51 SHORTCUTS

Why one tap goes far

03 / ENGINE
51
shortcuts
·
11
power macros
·
9
keys

The macros are written to trigger what Claude Code on Opus 4.8 does best.

Verification loops. “re-run until green”, “prove it passes”. Opus 4.8 scores 0% on uncritically reporting flawed results — “done” means done.
Dynamic workflows. Claude orchestrates tens to hundreds of parallel subagents, with iterative verification and state that resumes across interruptions.
Effort controls. Four levels — high by default, max when quality is the only variable.
Browser verification. The strongest browser agent tested (84% on Online-Mind2Web) — it can drive the UI and prove a change, not just write it.
Grounded, honest answers. “cite file:line”, “quote the error” — with 10× less overconfidence than Opus 4.7.
Guardrails. Destructive moves like force-push confirm first.