Multiplex AI agents in persistent terminal sessions
herdr is a background terminal server that keeps coding-agent sessions running through lid closes, network drops, and restarts, then reattaches.
0.8.0Add to Favorites
Why it matters
Run multiple AI agents simultaneously in a terminal multiplexer that persists sessions across disconnects and restarts, enabling agents to coordinate work through a socket API while providing real-time visibility into each agent's status and output.
Outcomes
What it gets done
Monitor multiple agent processes in split terminal panes with live status updates
Detach and reattach to running agent sessions from any terminal or SSH connection
Enable agents to spawn panes, read output, and coordinate through socket API
Extend agent workflows with plugins from the marketplace
Source
Get it from source
Spark does not host a copy of it.
Open sourceReports
Agent outcome reports
No reports yet
Overview
Herdr
herdr is a background server that hosts your coding-agent terminal sessions so they keep running through disconnects and restarts, marking each pane as working, blocked, or idle and exposing a socket API agents can drive themselves. Use it when running multiple coding agents in parallel and needing persistent, reattachable sessions with at-a-glance status; not needed for a single agent in one terminal you never detach from.
What it does
herdr describes itself as "the runtime your coding agents live on." It is a background server that hosts terminal panes rather than a terminal emulator itself: agent sessions keep running when you close the lid, drop the network, or restart the machine, and you reattach to the same panes afterward from any terminal, including over SSH. Every pane is marked working, blocked, or idle, so instead of scanning windows for the one that's stuck waiting on an answer, herdr flags it directly.
It is agent-native by design: the CLI and its socket API are the same surface agents themselves drive - an agent can spawn new panes, prompt another agent, or wait until a different agent is genuinely blocked before acting. herdr doesn't wrap or replace the coding agents it hosts (Claude Code, Codex, Cursor, opencode, Grok, and others); it owns their terminals rather than emulating them, so whatever CLI you already run keeps working as-is.
When to use - and when NOT to
Use it when you're running multiple coding agents in parallel and need sessions to survive disconnects, want to reattach to in-progress work from a different machine, or want a fast, at-a-glance read on which agent needs your attention versus which is still working. The interface supports both tmux-style prefix keys and direct mouse interaction (click, drag, split), so it fits either a keyboard-only or mouse-driven workflow.
Don't reach for it if you only ever run one agent in one terminal session with no need to detach and reattach - the value here is specifically in long-running, multi-pane, multi-agent sessions that need to persist independent of any one terminal window staying open.
Inputs and outputs
Inputs are the terminal commands and coding-agent CLIs you already run inside herdr's panes; herdr's own output is the pane state (working/blocked/idle) plus the socket API surface that agents use to spawn panes, send prompts to each other, or poll for another agent's blocked state. Sessions and their state persist across disconnects and machine restarts.
Integrations
herdr hosts existing coding-agent CLIs directly - Claude Code, Codex, Cursor, opencode, and Grok are named explicitly - without wrapping or replacing them. A plugin system extends panes and workflows beyond the built-in feature set, with a marketplace of available plugins. It ships as a single Rust binary with no Electron dependency, so it runs in whatever terminal you already use.
Who it's for
Developers running multiple coding agents at once who need sessions that outlive a closed laptop lid or a dropped connection, and who want a single place to see which agent is stuck and needs a response versus which is still working unattended.
herdr is licensed under the Apache License 2.0.
Source README
herdr
herdr.dev · install · quick start · docs
English · 简体中文
https://github.com/user-attachments/assets/043ec09f-4bdd-41d5-aee0-8fda6b83e267
the runtime your coding agents live on.
- always running - herdr is a background server; the terminals live inside it. close the lid, drop the network, restart the machine - agents keep working and sessions come back. reattach from any terminal, or over ssh.
- never hunt for the stuck one - every pane is marked working, blocked, or idle. when an agent stops and needs an answer, herdr says so.
- agent-native - the cli and socket api are the same surface agents drive: spawn panes, prompt each other, wait until another agent is genuinely blocked. agent skill →
- runs what you already run - claude code, codex, cursor, opencode, grok and the rest. herdr doesn't wrap or replace them, it just owns their terminals.
- keyboard and mouse, both first-class - tmux-style prefix keys and click, drag, split. pick per moment, not per tool.
- plugins - extend panes and workflows. browse the marketplace →
- one rust binary, no electron - runs in whatever terminal you already use.
install
curl -fsSL https://herdr.dev/install.sh | sh
or brew install herdr · mise use -g herdr · windows beta: powershell -ExecutionPolicy Bypass -c "irm https://herdr.dev/install.ps1 | iex" · binaries
then start it where the work lives:
herdr
run your agents, split panes, walk away. ctrl+b q detaches, herdr reattaches. quick start →
docs
everything lives at herdr.dev/docs: quick start · concepts · supported agents · keyboard · configuration · session state · remote · integrations · plugins · socket api
thanks
Terminal Trove and every backer listed in SPONSORS.md - thank you 🐑
enterprise / partnership: hey@herdr.dev
agent instructions
if you are an ai agent helping with this repository, read AGENTS.md before making changes and read CONTRIBUTING.md before opening issues or PRs.
development
git clone https://github.com/herdrdev/herdr
cd herdr
cargo build --release
just test # unit tests
just check # formatting, tests, and maintenance checks
FAQ
Common questions
Discussion
Questions & comments · 0
Sign In Sign in to leave a comment.
