Launching today
Agent terminal
One terminal. Every agent. Total clarity.
32 followers
One terminal. Every agent. Total clarity.
32 followers
A terminal built for AI agents. It comes with project workspaces, live process metrics, and native support for Claude Code and Codex. Check it out at github.com/DaniAkash/agent-terminal.
Hey Product Hunt 👋
Something @dani_akash_ built that I genuinely wish existed sooner: Agent Terminal.
If you've started running multiple AI coding agents at once — Claude Code in one tab, Codex in another, a dev server somewhere, a git push you forgot about — you already know the chaos. You're constantly tab-switching just to find which agent is waiting on you, and which one quietly gave up an hour ago.
Dani felt that too. So he built a terminal around one simple idea: you should always know what's happening, without having to go look.
A few things that make it different:
Tabs organized by project, not dumped into a flat list. Your brain thinks in projects. Now your terminal does too.
A glyph on every agent tab — idle, working, waiting on you, or done. No more clicking around to find the one that needs your attention.
A live status bar that shows memory, ports, git state, and which model's running — updated every 2 seconds, not whenever you remember to check.
Sessions that survive accidental closes. The PTY reconnects. That 3am cmd+W is no longer a disaster.
It's open source (MIT), macOS for now, built on Tauri + Rust — fast and genuinely out of your way.
Dani is reading every comment today. What part of your AI coding workflow still feels broken? Drop it below 👇
— Himanshu (helping out @dani_akash_)
This is exactly the kind of problem you only notice after working with multiple agents for a while 😅
Constant tab switching just to figure out what’s stuck is real. The status indicators and session recovery look super useful.
Gonna give this a try.
Would be great to see support for Antigravity as well.
@a_syed_abdul_rahman Unfortunately it's limited to terminal only agents for now... So we'll be looking into adding gemini cli support next!
apideck
Awesome work @himanshu512 and @dani_akash_
I want to know, does this provide notification/pings when Claude/Codex has completed a run and is idle?
@himanshu512 @srbhr yup notifications are supported for both!
apideck
Thanks, I think one amazing feature would be sandboxing so that AI Agents can control the entire terminal and perform actions, while we don't need to worry about what happened. Also, logs of each command being run. Live deterministic monitoring types.
@srbhr sandboxing is beyond the scope of this project, you might wanna try the daytona sdk...
Bullet.so - Build websites from Notion
Been looking for this. Definitely gonna try it. Quick question: Does it support claude multi-agent teams?
@aswin_kumar right now it just functions as a normal terminal. But I'm looking into the sub agents process to see if I can surface any useful information...
Writesonic
This solves a real problem. I've been living with n terminal tabs open, constantly clicking through them to figure out which agent is spinning and which one stopped silently. Congrats on shipping this — trying it tonight.
@sanujsood do share your experience with us 😁