DEV.co
AI Coding Agents · agent-of-empires

agent-of-empires

Agent of Empires is a Rust-based session manager for running multiple AI coding agents (Claude, Mistral, Gemini, etc.) in parallel, accessible via terminal UI or web dashboard. It handles agent lifecycle, git worktrees, Docker sandboxing, and cross-device access so developers can monitor and control distributed agents from laptop or mobile.

Source: GitHub — github.com/agent-of-empires/agent-of-empires
2.8k
GitHub stars
254
Forks
Rust
Primary language
MIT
License (OSI-approved)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

FieldValue
Repositoryagent-of-empires/agent-of-empires
Owneragent-of-empires
Primary languageRust
LicenseMIT — OSI-approved
Stars2.8k
Forks254
Open issues148
Latest releasev1.12.0 (2026-07-01)
Last updated2026-07-07
Sourcehttps://github.com/agent-of-empires/agent-of-empires

What agent-of-empires is

Built in Rust with tmux as the execution backbone, AoE provides a TUI (terminal UI) and PWA-capable web dashboard for orchestrating multiple LLM-based coding agents. Sessions persist independently of the client connection; the web API and Agent Client Protocol enable structured state rendering and remote orchestration.

Quickstart

Get the agent-of-empires source

Clone the repository and explore it locally.

terminalbash
git clone https://github.com/agent-of-empires/agent-of-empires.gitcd agent-of-empires# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

Parallel multi-agent development workflows

Launch many Claude Code or OpenCode instances across git branches or repos in one dashboard, each isolated in its own tmux session. Ideal for evaluating agent performance or running parallel code-generation tasks.

Remote/mobile agent monitoring

Expose sessions via web dashboard (PWA-installable) with Tailscale Funnel or Cloudflare Tunnel auth. Check on long-running agents, approve tool calls, and review diffs from a phone or tablet without SSH.

Sandboxed agent execution with session persistence

Run agents in Docker containers with shared auth volumes, and resume multi-turn conversations across reboots via Claude session storage. Useful for untrusted or experimental agent prompts.

Implementation considerations

  • Requires tmux as a hard dependency; Docker/Podman optional but recommended for sandboxing. Verify tmux version compatibility for your OS.
  • Web dashboard requires starting `aoe serve`; remote access needs Tailscale Funnel, Cloudflare Tunnel, or equivalent. Plan network/auth topology upfront.
  • Git worktree support assumes familiarity with git workflows; misconfigured worktrees can lead to branch conflicts. Test on a staging repo first.
  • Docker sandboxing requires volume mounts for auth credentials; ensure your agent commands export credentials safely or use mounted secret files.
  • Session resume for Claude depends on storing conversation state; verify your session storage backend (local file, cloud) is backed up.

When to avoid it — and what to weigh

  • Simple single-agent workflows — If you only run one Claude Code instance at a time, the overhead of tmux + AoE adds friction. Native Claude IDE or cursor may be simpler.
  • Windows-only deployment — AoE targets Linux and macOS. Windows is not mentioned as supported; requires WSL or alternative tooling.
  • Requirement for agent-agnostic state serialization — AoE is tightly coupled to tmux sessions and Agent Client Protocol. If you need deterministic, portable agent state snapshots across different backends, this may not provide that out of the box.
  • Production multi-tenant SaaS orchestration — AoE is designed for personal or small-team agent management. No mention of audit logging, RBAC, or compliance features needed for regulated multi-tenant use.

License & commercial use

MIT License. Permissive open-source license permitting unrestricted use, modification, and distribution for any purpose, commercial or otherwise, with attribution.

MIT is a permissive OSI license. Commercial deployment and modification are allowed. No license restrictions on commercial use, but review any proprietary extensions or vendor-specific agent integrations (Claude, Mistral, Gemini) for their own terms.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

SignalAssessment
MaintenanceActive
DocumentationStrong
License clarityClear
Deployment complexityModerate
DEV.co fitGood
Assessment confidenceHigh
Security considerations

Agent sessions run as the invoking user in tmux; no privilege isolation by default (Docker sandboxing optional). Web dashboard auth is optional (passphrase + QR via Tailscale Funnel or Cloudflare Tunnel); no built-in RBAC or audit logging. Verify credential handling when agents call external APIs; mounted auth volumes in Docker require careful permissions. No security audit or CVE history provided.

Alternatives to consider

Cursor IDE / VS Code Agent extensions

Native IDE integration for single-agent workflows; no tmux overhead. Lacks multi-agent parallelism and cross-device mobile access.

Anthropic's Claude API + custom orchestrator (Python/Node.js)

Full programmatic control; can implement custom session persistence, RBAC, and audit logging. Requires more engineering; no off-the-shelf TUI or web dashboard.

Langchain / LlamaIndex agent frameworks

Language-agnostic, extensible frameworks for multi-step agent workflows. Lack built-in session management, tmux integration, or mobile dashboards; require custom UI.

Software development agency

Build on agent-of-empires with DEV.co software developers

Install AoE to run multiple coding agents in parallel, monitor them from any browser, and keep sessions alive across reboots—all from a single dashboard.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

Related on DEV.co

Explore the category and the services that help you build with it.

agent-of-empires FAQ

Do agent sessions persist if I close the AoE TUI?
Yes. Each agent runs in its own tmux session in the background. Closing AoE does not kill sessions; reopen AoE and sessions are exactly where you left them.
Can I access agents from my phone?
Yes. Press `R` in the TUI to start the web dashboard server; it exposes a browser-accessible PWA with passphrase auth via Tailscale Funnel or Cloudflare Tunnel.
What happens if an agent makes destructive changes to my repo?
Each agent runs in its own git worktree or branch, isolating changes. Docker sandboxing (optional) adds container-level isolation. Always review diffs via the built-in diff view before merging.
Is AoE available for Windows?
Not explicitly supported. Linux and macOS are documented. Windows users would need WSL2 or a native port (requires community contribution).

Software developers & web developers for hire

DEV.co helps companies turn open-source tools like agent-of-empires into production software. Our software development services cover the full lifecycle — architecture, web development, integration, and maintenance — delivered by software developers and web developers who ship. Engage our software development agency to implement or customize it for your ai coding agents stack.

Coordinate your AI agents.

Install AoE to run multiple coding agents in parallel, monitor them from any browser, and keep sessions alive across reboots—all from a single dashboard.