Everything you need to know about the agent multiplexer
Find answers to common questions about herdr, installation, usage, and integration.
Comprehensive answers to help you get the most out of herdr. Click on any question to reveal its answer.
Installation and first steps
How to use herdr effectively
Supported AI coding assistants
Advanced capabilities
herdr is an agent multiplexer that lives in your terminal. It allows you to run all your coding agents in one terminal, providing a unified view of their status and activities. Each agent gets its own real terminal, not an app's imitation, so even full-screen TUIs render correctly.
herdr has multiple installation options:
Shell Install:
curl -fsSL https://herdr.dev/install.sh | sh
Windows Beta:
powershell -ExecutionPolicy Bypass -c "irm https://herdr.dev/install.ps1 | iex"
Homebrew:
brew install herdr
Nix:
nix run github:ogulcancelik/herdr
You can also download stable Linux/macOS binaries from the releases page.
herdr is a single ~10MB Rust binary that runs on:
It has no dependencies and runs inside your existing terminal.
Simply run herdr in your terminal. This will start or attach to a background server and open a workspace where you can run your agents.
herdr is mouse-native, so you can click and drag panes, tabs, and split borders. For keyboard users, ctrl+b is the prefix key for shortcuts.
Press ctrl+b then release, then press the action key:
shift+n - new workspacev or minus - split panesc - new tabw - switch workspacesq - detach (agents keep running)? - show all bindingsSee the keyboard guide for more details.
Yes! Run herdr on a VPS and access it from your local terminal using herdr --remote. This makes your local terminal the client of the remote server, preserving features like image pasting that break with plain SSH + tmux.
herdr --remote workbox
herdr --remote ssh://you@yourserver:2222
See the persistence and remote docs for more details.
herdr supports many popular coding agents with built-in detection:
Detection works out of the box with process-name matching and terminal-output heuristics. You can install official integrations for native session restore and semantic state reporting with herdr integration install .
herdr uses a combination of process-name matching and terminal-output heuristics to detect agent states:
The sidebar shows all agents with their current status at a glance.
Agents can use herdr's local Unix socket to create workspaces, split panes, spawn helpers, read output, and subscribe to state changes. Install the reusable skill with:
npx skills add ogulcancelik/herdr --skill herdr -g
Start with the agent skill docs and socket API docs.
While tmux provides persistence and panes, it was built before agents existed and has no awareness of agent states. herdr gives you:
See the full comparison with tmux, zellij, cmux, warp, conductor, and more.
When you detach (using ctrl+b q), a background server keeps all panes and agents alive. Nothing dies! You can reattach from any terminal, including your phone over SSH, and continue exactly where you left off.
Yes! herdr is dual-licensed:
Contact hey@herdr.dev for commercial licensing inquiries.
We welcome contributions! Check out the GitHub repository for:
herdr is built full-time, in the open, with no revenue behind it. You can support development by:
Thank you for your support! 🐑
Join thousands of developers who are already using herdr to manage their coding agents.