Every engine on one board
ChatGPT, Claude, Gemini, Grok, DeepSeek — or MegaPad hosted models with zero API setup. Six lanes, live status, one keyboard.
The command center for agentic work
OpenAI just shipped a $230 macropad that commands one model. MegaPad commands them all — ChatGPT, Claude, Gemini, Grok, DeepSeek — from an app on your desk, starting at free.
$99 refundable is live cash today (Pro + pad queue). Free download when repo/public host is ready · one-tap pay page
Why MegaPad
The big labs proved the category: agent keys, status lights, an effort dial, one surface to command it all. They built it for their model, their subscribers, in a limited run. MegaPad is that command center as software — for every model, for everyone, today.
ChatGPT, Claude, Gemini, Grok, DeepSeek — or MegaPad hosted models with zero API setup. Six lanes, live status, one keyboard.
Describe the job — slides, a website plan, a campaign, an email sequence. An agent runs it while you do something else. You review the result and accept or reject it.
MegaPad runs on your machine. Your API keys stay on your disk. Prefer no keys at all? Run hosted models through your MegaPad account instead.
Engines
Point two engines at the same problem. Keep the better answer. That's not possible on a single-model device — it's the default here.
Built for work, not for prompts
Pick a job from the board and an agent takes it end to end. No API consoles, no prompt engineering, no terminal. When it's done you get one decision: accept or reject.
The honest comparison
Respect where it's due: agent keys, status lights and a reasoning dial are the right idea. Here's the difference.
| Codex Micro | MegaPad | |
|---|---|---|
| Engines | Codex only | ChatGPT, Claude, Gemini, Grok, DeepSeek + hosted |
| Price to start | $230 | Free |
| Availability | Limited run | Download — software ships continuously |
| Who it's for | Developers running coding agents | Anyone with work to hand off |
| Agent status keys | Yes — hardware LEDs | Yes — six lanes on screen; hardware later, same protocol |
| Reasoning / effort dial | Yes | Yes |
| Compare two models on one task | — | Yes — accept the better answer |
| Where your keys live | ChatGPT account | Your machine (or no keys via hosted) |
Codex Micro is an OpenAI × Work Louder product. Comparison based on public launch coverage, July 2026. No affiliation.
How you get it
MegaPad is a desktop app with a local engine room. Mac today, Windows build in progress. The physical pad comes later on the same protocol — reserve it if you want the desk ritual.
Mac app now. Windows next. One window, no browser tabs.
Sign in for hosted models — or paste your own keys, stored locally. ChatGPT sign-in supported.
Six agent lanes. Live status. Accept what ships, reject the rest.
Pricing
Cash live now
$99
deposit today, $299 total, refundable
$0
App when download is public
Software only
$149
Early lifetime, no pad
Founding is Pro software plus a refundable place in the Edition One hardware queue (target window: 2027). The pad below is a design draft, not a shipping product. We say so out loud, unlike some.
One email when the download opens. Founding deposits are live now.