Agent multiplexer · lives in your terminal

One terminal. The whole herd.

Herdr is to coding agents what tmux is to terminals: a multiplexer that runs where your agents run — your machine, your server, anywhere you can ssh. See blocked, working, and done at a glance, click anything, close the laptop. Nothing dies.

Stable install
$ curl -fsSL https://herdr.dev/install.sh | sh
Windows beta · preview only Preview docs
PS> irm https://herdr.dev/install.ps1 | iex

Stable Linux/macOS · Windows preview beta · no Electron

The whole herd in one terminal. Click the sidebar — it's a real layout.

# used in the wild

Popular with engineers from

JetBrainsJetBrains
DockerDocker
VercelVercel
KimiKimi
SentrySentry
GoogleGoogle
NVIDIANVIDIA
AWSAWS
ByteDanceByteDance
TencentTencent
AlibabaAlibaba
SalesforceSalesforce
IBMIBM
AtlassianAtlassian
WhopWhop
AutomatticAutomattic

# not an app

Runs where your agents run.

Desktop agent managers are stuck on the machine with the GUI. Herdr is a client–server multiplexer: agents live on whichever box does the work, and you attach from anywhere — your laptop, any SSH client, even a phone.

local

Your machine

Start a local session, split panes, create tabs, and keep agents running while your terminal comes and goes.

$ herdr
ssh

Any SSH client

Use Herdr like tmux on a remote box. The session stays on the server after you detach.

$ ssh you@server
$ herdr
thin client

Remote attach

Point your local Herdr client at an SSH config host and stream the remote TUI back to your terminal.

$ herdr --remote workbox

# responsive tui

Mobile-first when the terminal gets small.

Herdr stays usable over SSH from a phone or tablet. The terminal view remains real, while narrow screens get a switcher built for touch-sized decisions.

Herdr agent session over SSH on a phone
agent session over SSH
Herdr responsive switch menu on a phone
responsive switch menu

screenshots taken with moshi ❤️ on iPhone

# the missing layer

Keep your terminal. Add the control surface.

Desktop agent terminals give you app UI. tmux gives you persistence. Herdr sits between them: a lightweight TUI process with real terminal panes, clickable layout, agent state, detach/reattach, remote attach, and an API your agents can drive.

not a terminal emulator

Ghostty, Kitty, iTerm, Alacritty — your terminal stays.

not a browser dashboard

No web view, no account, no hosted control plane.

more than tmux for agents

Persistence plus clickable panes, agent state, an API.

# positioning

Most agent managers are apps. Herdr is a multiplexer.

tmux and Zellij own persistent terminal sessions but don't understand agents. Desktop agent apps understand agents but live on one machine. Herdr keeps the multiplexer in your terminal and makes it agent-aware. See the full comparison →

Capability tmux / Zellij agent apps worktree orchestrators Herdr
Runs inside your terminal
Persistent PTY sessions limited embedded
Remote SSH attach limited remote projects
Semantic agent state partial workspace status
Direct agent attach
Agents can orchestrate it scriptable partial workflow-owned

Runs inside your terminal

tmux/Zellij agent apps worktree apps Herdr

Persistent PTY sessions

tmux/Zellij agent apps limited worktree apps embedded Herdr

Remote SSH attach

tmux/Zellij agent apps limited worktree apps remote projects Herdr

Semantic agent state

tmux/Zellij agent apps partial worktree apps workspace status Herdr

Direct agent attach

tmux/Zellij agent apps worktree apps Herdr

Agents can orchestrate it

tmux/Zellij scriptable agent apps partial worktree apps workflow-owned Herdr

# your terminal

Same shell, same keybinds, real panes.

No rebuilt chat view. Real processes in real PTYs, with layout, clicks, state, and persistence added around them.

  • Use the mouse for panes, tabs, workspaces, and menus.
  • Keep your terminal fonts, shortcuts, SSH, and shell setup.
  • Stay keyboard-first, or disable mouse input entirely.

# agents can drive it

The UI is also a control surface.

A CLI and JSON socket API expose workspaces, panes, output, and waits. Your agents can run their own workspace.

# create workspace structure
herdr workspace create --cwd ~/project --label api
herdr tab create --label logs

# split a pane and run work
herdr pane split 1-1 --direction right
herdr pane run 1-2 "just test"

# wait, inspect, and continue
herdr wait agent-status 1-1 --status done
herdr pane read 1-2 --source recent-unwrapped

# agent awareness

Keep the herd in sight.

Workspace rollups

Each workspace rolls up to its most urgent agent state, so the whole session is scannable.

Agent list

See which agents are blocked, working, done, idle, or unknown without opening every pane.

blocked working done idle

# integrations

Bring your existing agents into the herd.

Pi
Claude Code
Codex
Droid
Amp
OpenCode
Grok CLI
Hermes
Cursor
Antigravity
Kimi
Kiro
Copilot
Qoder CLI

Any terminal agent works out of the box. Integrations add richer state and session resume.

# start the herd

Install Herdr, run two agents, detach once.

The value shows up the first time the herd keeps running after your terminal disappears.

light [/]