█

AI/TLDR

herdr

A terminal multiplexer rebuilt around long-running coding agents instead of shell sessions

GitHub★ 39kWebsiteherdr.dev
Coding-Agent Harnesses & RuntimesOpen source
Language
Rust
License
Apache-2.0
$curl -fsSL https://herdr.dev/install.sh | sh

Overview

herdr describes itself as "the runtime your coding agents live on". It is a terminal application, written in Rust and shipped as a single binary with no Electron runtime, that takes the multiplexer idea — panes, layouts, detach and reattach — and reshapes it for work where the thing occupying a pane is an agent rather than a person at a prompt.

The practical difference shows up in three places. Terminals keep running on a background server when the client disconnects, and layouts restore after a restart (the original processes themselves do not survive a reboot). Local work and remote SSH machines are consolidated into one interface, so a fleet of agents spread over several hosts is visible in a single window. And every pane carries a status — working, blocked or idle — which is the piece a normal multiplexer cannot give you, because "blocked waiting for approval" is not something a shell reports.

Agents are treated as first-class clients, not just processes being watched. They talk to herdr through a CLI and a socket API, which lets an agent open its own panes and coordinate with other agents running alongside it. It works with the mainstream coding agents — Claude Code, Codex, Cursor, OpenCode and Grok among them — and is driven either by tmux-style prefix-key bindings or with the mouse. A plugin system with a marketplace covers extensions.

What it does

  • Session persistence: terminals keep running on a background server after the client disconnects, and layouts restore on restart
  • Local machines and remote SSH hosts consolidated into a single interface
  • Per-pane agent status — working, blocked or idle — visible at a glance
  • Agent-native CLI and socket API: agents create their own panes and coordinate with each other
  • Works with Claude Code, Codex, Cursor, OpenCode, Grok and similar tools
  • tmux-style prefix-key bindings plus mouse interaction
  • Single Rust binary, no Electron; plugin system with a marketplace

Getting started

herdr installs as one binary. The commands below come from the project README.

Install

The install script covers macOS and Linux.

bashbash
curl -fsSL https://herdr.dev/install.sh | sh

Or install with a package manager

Homebrew and mise are both supported; a PowerShell install script is published for Windows, and binary releases are available for several architectures.

bashbash
brew install herdr
# or
mise use -g herdr

Start it

Running the binary with no arguments opens the multiplexer.

bashbash
herdr

Detach and reattach

Detaching leaves the server — and the agents in it — running. Re-running `herdr` reattaches to the same layout.

bashbash
ctrl+b q   # detach
herdr      # reattach

Run your agents in panes

Launch a coding agent inside a pane as you normally would; herdr tracks whether it is working, blocked or idle, and the agent can open further panes through the CLI and socket API.

bashbash
claude
# or codex, opencode, cursor-agent, …

Commands and code are distilled from the project's own documentation — always check the official repo for the latest.

When to use it

  • Reach for it when several coding agents run at once and you need to see which ones are blocked on you
  • Reach for it when agent work spans local and remote SSH machines and you want one window over all of them
  • Reach for it when long-running agent sessions must survive a disconnect or a client restart
  • Reach for it when you want agents to open and coordinate their own panes rather than driving a multiplexer by hand

How herdr compares

herdr alongside other open-source coding-agent harnesses & runtimes tools AI/TLDR tracks, ranked by GitHub stars.

ToolStarsWhat it does
CC Switch★ 133kA cross-platform desktop app that manages provider configuration for eight coding-agent CLIs from one place, with a local failover proxy, MCP and skill management, and spend tracking.
Oh My OpenAgent★ 69.1kOpinionated agent bundle for OpenCode and Codex CLI that ships preconfigured sub-agents, lifecycle hooks and built-in MCP servers behind a single ultrawork command.
oh-my-claudecode★ 39.2kA plugin and CLI that adds multi-agent orchestration to Claude Code: staged autopilot workflows, parallel tmux workers, and advisors that route a second opinion to Codex, Gemini, Grok or Cursor.
herdr★ 39kA terminal multiplexer rebuilt around long-running coding agents instead of shell sessions
AionUi★ 32.9kAn open-source desktop workspace that pairs a built-in agent engine with dozens of external CLI agents, adding scheduled runs, MCP tools and remote access from a WebUI or chat apps.
cmux★ 27.2kA Ghostty-based macOS terminal built around coding agents: vertical tabs that show git branch, PR status and ports, attention notifications when an agent needs you, and a scriptable built-in browser.
cc-connect★ 15.5kA Go daemon that bridges local coding agents — Claude Code, Codex, Cursor Agent, Gemini CLI, OpenCode, Kimi CLI and more — to Feishu, DingTalk, Slack, Telegram, Discord, LINE and WeCom, so you drive them from chat.
Claude Code Haha★ 14.5kA cross-platform desktop workspace for Claude Code: multi-session workspaces, worktree launch, diff review, a built-in browser preview, GUI permission approval, per-session model choice and visual MCP and subagent managers.