If you juggle several AI coding agents at once, you know the window shuffle: one IDE here, another terminal there, and a browser full of issues, pull requests, docs and localhost previews. A new open-source tool called tabsh flips that setup around — instead of bringing the browser into your IDE, it brings your coding sessions into a browser tab.
tabsh (Credits: Product Hunt)
The idea is simple. Most agentic development environments (ADEs) try to pull everything into one big editor. tabsh takes the opposite approach: since your work already happens in the browser, why not manage your agents there too? A small Rust daemon puts all your sessions, across all your projects, onto a single Kanban board or list in a browser tab.
tabsh: Work With AI Agents With an IDE in Your Browser Tab
Each session on the board shows its status, updated automatically through agent hooks — including when an agent needs your input. That last detail matters more than it sounds. When you run several agents in parallel, the bottleneck is usually not the agents themselves; it is noticing which one is stuck waiting for you. tabsh surfaces that at a glance, so you can unblock the right session instead of cycling through windows.
You can work in real terminals or chat panels, with support for Claude Code, Codex, Cursor CLI, OpenCode and more. The point is not to replace your editor, but to give parallel sessions a single control surface. One board, every project, every agent, with its current state visible.
How tabsh Works — and What It Costs
Everything runs locally. There is no account and no remote server, according to its Product Hunt listing, which should appeal to developers who are wary of sending their codebase or session data to yet another cloud service. It runs on macOS and Linux, and it is free and MIT-licensed — meaning you can use it, fork it and adapt it to your own workflow.
The maker’s own reasoning is relatable: their ADE was the last thing they kept switching to, because issues, pull requests, documentation, design work and localhost previews all already lived in the browser. Moving agent management into that same place removes one more context switch from a workflow that already has too many.
tabsh will not be for everyone. If you run a single agent in a single project, a board adds little. But if you regularly spin up several agents across several repositories — one fixing a bug, another writing tests, a third reviewing a pull request — a single view of what each one is doing, and which one needs you, could save real time and frustration.
Also Read: 3 Open-Source Mac Tools You Should Install Today: Maccy, LuLu and LocalSend and I Stopped Uploading Private Documents To AI — This Free App Handles Them On My PC With Ollama
Source: Product Hunt — read the full story there.
