✉️ Email · $30/mo

Can I vibe code Superhuman?

The fastest email client: shortcuts, splits, AI triage.

🟡Kinda — with gapsest. a weekend

A keyboard-first Gmail client is a real but rewarding project: the Gmail API gives you threads, labels, and send; the work is in the interaction quality — instant-feeling optimistic UI, command palette, split inboxes. You can get surprisingly far in a weekend and genuinely enjoy the result. What you won't replicate solo is the years of latency obsession and deliverability-adjacent features. At $30/month, the DIY math is tempting.

The one-shot prompt
Build a keyboard-first Gmail client for my account. OAuth with the Gmail API (read, modify, send). UI: a three-pane inbox (list, thread, context), j/k to move, e to archive, r to reply inline, cmd+k command palette for every action, and / for search using Gmail query syntax. Add a split inbox: rules stored in my database classify threads into Important and Everything Else (sender domains, keywords), with a one-key move-and-teach action. Include snippets (;sig etc.) and send-later via scheduled function. Optimistic UI updates with rollback on API failure. Next.js, deployable to Vercel.
Paste into Claude Code, Codex, or your agent of choice — then steer. Learn the moves in Prompt Engineering.
What you build

A Gmail client with a command palette, j/k navigation, e to archive, split inbox (important vs rest via rules), snippets, and send-later.

What you give up
  • Sub-100ms feel everywhere — their whole engineering culture is latency
  • AI triage/drafting tuned on email specifically
  • Read statuses, follow-up reminders, and calendar overlay polish
  • Multi-account and offline behavior that never breaks
Why people still pay

Power emailers process hundreds a day; if it saves 20 minutes daily, $1/day is nothing. They're buying speed as a feeling, which is brutally hard to clone.

polishlatency engineering
Related
Actually build it

A prompt is the first move, not the whole game. Course 01 teaches you to take a prompt like this one to a shipped, working app — reviewing, correcting, and steering the agent the whole way.

Start Course 01 →