<2-3 sentences summarising what the project addresses and solves, the target audience, and the benefit.> herdr is a terminal-native workspace manager designed to supervise multiple coding agents simultaneously within a single terminal window. It targets developers and power users who run various AI coding assistants and need a unified interface to manage their parallel execution, monitor their status, and facilitate inter-agent communication. herdr enhances productivity by providing context-aware workspaces, tabs, and panes, eliminating the need to switch between multiple terminals or GUIs.

How It Works

herdr operates as a single Rust binary, embedding itself within existing terminal emulators like Alacritty, Kitty, or even tmux. Its core innovation is a 'supervision layer' that adds awareness and management capabilities to terminal-based coding agents. It organizes work into workspaces, tabs, and panes, automatically detecting running agents by inspecting foreground processes and terminal output. Agents can communicate and coordinate via a local socket API, enabling complex workflows where agents can spawn new panes, wait for each other, and exchange data.

Quick Start & Requirements

• Install: curl -fsSL https://herdr.dev/install.sh | sh
• Prerequisites: Linux or macOS.
• Resources: Official documentation, quick start guides, and integration details are available at herdr.dev.

Highlighted Details

• Terminal-native workspace management with workspaces, tabs, and panes, offering a lightweight, project-context-first approach.
• Automatic agent detection via process inspection and terminal output heuristics for supported agents (pi, Claude Code, Codex, Droid, Amp, OpenCode) and custom integrations.
• Visual status indicators (🔴 blocked, 🔵 done, 🟡 working, 🟢 idle) in a sidebar, with configurable toast notifications or sounds for urgent events.
• Agent-to-agent communication via a local socket API for creating workspaces/tabs/panes, reading output, and orchestrating workflows.
• Supports multiple terminal emulators (ghostty, alacritty, kitty, wezterm, tmux) and offers 9 built-in themes with customization options.

Maintenance & Community

The project was notably built almost entirely through AI coding agents, with human supervision. No explicit community links (Discord/Slack) or contributor/sponsorship details are provided in the README.

Licensing & Compatibility

• License: AGPL-3.0.
• Compatibility: This strong copyleft license requires any distributed modified versions to also be open-sourced under AGPL-3.0, potentially restricting use in closed-source commercial products without open-sourcing those products.

Limitations & Caveats

The project's AI-driven development suggests potential for "rough edges." The AGPL-3.0 license imposes significant obligations for distribution, requiring derived works to also be open-sourced under AGPL-3.0. Automatic agent detection relies on heuristics and process inspection, which might not be foolproof for all agents or configurations.