tterm

A terminal, a real browser, and Claude Code under one roof

Productivity

Artificial Intelligence

Vibe coding

Visit WebsiteSee on Product Hunt

Cool product! I’ve actually been looking for something like this. Out of curiosity, what would you say are the main differences between this and tools like cmux, Orca, or Muxy?

Fewer than you'd expect, at least for our shape of work. A cycle here is roughly 40 reads and 5 writes, so the gate fires a handful of times a session and everything else stays fully autonomous. There's no model in the path either, it's a selector and URL allowlist, so the token cost is zero. What actually killed the annoyance was making approval sticky per domain for the session: I confirm the first push and not the next four.

the browser-session sharing thread here is a great read. curious about the other direction: if I've got two or three project rows stacked for different repos, is each row's Claude pane fully isolated from the others, or could a prompt in one row's context ever end up seeing something from a different project's files or conversation history?

@tanay good to know a follow-up prompt usually gets you unstuck. checkpointing would be a nice safety net for the cases where the rebuild goes sideways in a way that's hard to describe back to the model though - worth adding even as a manual 'save state' button before you ask it to touch the cockpit itself.

Ha, I've seen that exact behavior, the model bailing right before the irreversible step and going 'you take it from here.' It's a nice instinct but I wouldn't build a guarantee on it, since the same model that stops 90% of the time will confidently push the button the other 10. What worked for us was moving the decision out of the model entirely: the confirm gate fires on the action type, any write, delete, or spend, so it's the same whether Claude flagged it or not. Then the model bailing early is a bonus, not the safety net.

Appreciate you owning the tradeoff so plainly, most people would hand-wave it. Since you're clearly keeping the signed-in mode (I would too, that's where the useful work is), the thing that softened it for us was splitting read from write: the agent browses freely, but anything that changes state, a push, a delete, a purchase, waits on a one-tap confirm. Keeps most of the flow autonomous while the irreversible part stays yours. Would that fit tterm's diff-review philosophy, given you already gate commits hunk by hunk?

The layout is genuinely nice, basically treating each project like a little row you can stack while juggling multiple. Kind of wild that you can ask Claude in the pane to add something and the app just rebuilds itself around you.

the self-rebuilding feature is wild, asked the claude pane for a dark mode toggle and watched it hot reload right in front of me. also appreciate the no-account stance, just downloaded and started using it.

The bundled Chromium carrying your Chrome cookies is the detail I keep circling. I run agents against logged-in Chromium too, and each one gets its own throwaway profile on purpose: if the Claude pane can drive a browser already logged into your email or GitHub, a wrong click there executes as you, against your real accounts. That is a much bigger blast radius than a bad hot-reload. Is the in-app browser sharing your everyday Chrome session, or is it a separate profile the agent is scoped to?

No account and no telemetry is the right call, but the embedded Chromium carrying my Chrome bookmarks and cookies is the part I would want scoped before running this daily. Does the browser pane read my live Chrome profile directory, or a copy made at first launch? And since the Claude pane sits in the same app, can the agent drive that browser against sessions I am already authenticated into, or is agent-initiated navigation confined to a separate profile or an origin allowlist?

The self-rebuilding feature is genuinely wild. One thing I'd love: a way to bookmark or save the exact stack layout across projects so I can pick up right where I left off when I jump back in.

Love the no-editor, diff hunk by hunk flow, that's a clever take on reviewing AI generated code. One thing I'd find useful is a way to save session state per project so when you reopen a workspace the explorer tree, open files, and even scroll position come back exactly where you left them. Right now it sounds like every launch starts from scratch and that would save a lot of clicking.

stackable project rows with the explorer, Claude, and diff viewer side by side genuinely sped up how I review changes across multiple repos at once. Asking Claude for a tweak and watching the app hot-reload with it still feels a bit like magic.

finally a mac app that gets out of the way for claude code, and the hunk-by-hunk diff review is genuinely faster than clicking through vscode. weirdly charmed by the self-rebuild feature too

Finally a Claude wrapper that doesn't try to be an editor. Stacking multiple projects side by side with the diff review built in is exactly what I needed when juggling a few repos at once.

as a real Chromium browser is bundled in, would love to see a way to save reusable diff-view layouts or presets per project type, makes flipping between code reviews and Claude chats way faster

The hot-reload from asking Claude for a feature is genuinely clever, and committing without ever opening an editor is a real UX call, not just an omission.