agent-of-empires
agent-of-empires is an open-source terminal session manager designed specifically for AI coding agents on Linux and macOS. It provides a centralized interface to launch, monitor, and control multiple AI-powered coding sessions simultaneously, making it easy to run agents in parallel across different projects and tasks. The tool integrates with popular AI coding assistants like Claude Code, OpenHands, and others, wrapping their terminal interfaces into managed sessions that persist across disconnections. With agent-of-empires, users can group sessions by project directories, view real-time terminal output, and manage agent lifecycles. It features a workspace-centric web dashboard that organizes sessions in a sidebar, displays terminal panes, and even offers inline diffs for code changes. The tool supports automatic reconnection, so if a network drops, it retries and restores the session seamlessly. Sessions can be started, stopped, and monitored from both the command line and the web UI. Under the hood, agent-of-empires is built in Rust for performance and reliability. It uses PTY (pseudo-terminal) handling to spawn and manage agent processes, and includes a built-in web server for the dashboard, which can be accessed locally or over a network. The tool is designed for developers and teams who heavily use AI coding agents, enabling efficient multitasking and project management. Additionally, agent-of-empires is open source with an active community, as evidenced by its GitHub repository with thousands of stars and frequent commits. It aims to be the central hub for orchestrating AI coding work, reducing context switching and improving productivity when working with multiple AI assistants.
- Category
- AI Agents
- Pricing
- Open Source
Problem solved
Developers and engineers using AI coding agents often need to run multiple agent sessions in parallel, each working on different tasks or projects. Switching between terminals, keeping sessions alive, and monitoring progress becomes chaotic. agent-of-empires solves this by providing a unified session manager that persists terminal sessions, allows parallel execution, and offers a dashboard for oversight, so users can efficiently manage many concurrent AI agents without losing context or state.
Use cases
- Running multiple Claude Code sessions in parallel to tackle different parts of a codebase.
- Persisting long-running AI agent sessions across SSH disconnections on remote servers.
- Organizing AI coding tasks by project workspace, with each workspace containing multiple agent sessions.
- Monitoring agent progress and reviewing code diffs in real-time via the web dashboard.
- Quickly spinning up agent sessions for ad-hoc code reviews or bug fixes without manual terminal management.
Audience
- Developers
- ML Engineers
- Data Scientists
- Researchers
Pros
- Open-source and free to use.
- Lightweight and fast, built in Rust.
- Supports simultaneous parallel agent sessions.
- Web dashboard for visual management and diff viewing.
- Session persistence with auto-reconnection.
Cons
- Primarily terminal-based, which may deter users who prefer GUIs.
- Currently only supported on Linux and macOS, not Windows.
- Requires familiarity with AI coding agents and terminal multiplexing concepts.
Official website: https://github.com/njbrake/agent-of-empires
Canonical LambdaBase page: https://www.lambdabase.com/tools/agent-of-empires