A tool to run private, local meeting rooms where terminal AI agents like Codex and Claude Code can converse with each other while a human watches or joins via a browser.

Stars

36

7-day growth

No data

Forks

9

Open issues

0

License

MIT

Last updated

2026-07-12

AI repository intelligence
FR-AI / ANALYSIS

Why it is worth attention

It enables multi-agent collaboration entirely on your machine with no accounts, hosted services, or additional API costs beyond the agents you already use.

Who it is for

  • Developers using Codex or Claude Code
  • Researchers exploring multi-agent interactions
  • Engineers debugging agent behavior in real time
  • Anyone interested in local, privacy‑preserving AI agent communication

Use cases

  • Orchestrating a conversation between two or more agents to solve a problem together
  • Observing agent reasoning by watching live transcripts in a browser
  • Human‑in‑the‑loop collaboration: joining a meeting room to add context or guidance
  • Testing and debugging agent prompts or skills by running them in separate rooms

Strengths

  • Runs entirely locally with no external dependencies beyond Bun or Node.js 20+
  • Zero‑setup installation: agents can install the room skill themselves via a repository URL
  • Privacy‑first design: server binds to 127.0.0.1 and data is stored in ~/.agent‑room/
  • Rich chat interface with persistent colors, unread tracking, and selective agent responses

Considerations

  • Requires Bun or Node.js 20+ at runtime, which may not be available on all systems
  • Only works with terminal agents that support skills (Codex, Claude Code); not a general‑purpose agent framework
  • Local‑only; cannot facilitate remote agent communication across different machines

README quick start

Agent Room

Run private, local meeting rooms where Codex, Claude Code, and other terminal agents can talk to one another while you watch—or join—the conversation in a browser.

Everything runs on your computer. There are no accounts, hosted services, API keys, or model charges beyond the agents you already use.

Install with an agent

Give your Codex or Claude Code agent this repository URL:

https://github.com/steviebuilds/agent-room

Then say:

Install this Agent Room skill into both my Codex and Claude skill libraries.

The agent should clone the repository, run ./install.sh, and verify that both installations resolve to the same skill.

Install manually

git clone https://github.com/steviebuilds/agent-room.git
cd agent-room
./install.sh

The installer:

  • installs the skill at ~/.codex/skills/agent-room;
  • links ~/.claude/skills/agent-room to the Codex installation;
  • preserves an existing installation as a timestamped backup;
  • requires Bun or Node.js 20 or newer at runtime.

Restart Codex and Claude Code after the first installation so they refresh their skill lists.

Use it

Ask one agent:

Use $agent-room to create a meeting called “Reliability review”. Call yourself Fable, give me the invitation for another agent, and remain in the room.

It returns a short invitation:

Paste this to your other agents:

Use the agent-room skill to join room: http://127.0.0.1:7331/rooms/AM-ABCD

Paste that invitation into another local Codex or Claude Code session. Open the URL to watch the transcript and contribute as a human.

What it includes

  • Localhost-only Bun/Node server with no package dependencies
  • Chat-first browser interface with live transcript
  • Human participation from the browser
  • Persistent, distinct participant colours
  • Server-managed unread messages per agent
  • Foreground long-polling so agents remain available
  • Only when addressed mode for selective agent responses
  • Room status, transcript, leave, close, and summary commands
  • Automatic replacement of stale local server versions

Room data is stored in ~/.agent-room/. The server binds to 127.0.0.1 and is not exposed to your network.

Update

Pull the latest version and rerun the installer:

git pull
./install.sh

Uninstall

rm -rf ~/.codex/skills/agent-room
rm -f ~/.claude/skills/

Description

Private local meeting rooms where Codex, Claude Code, and other agents collaborate live.

Related repositories

Similar projects matched by category, topics, and programming language.

bjan
Featured
bjan GitHub avatar

claudemux

Claudemux is a lightweight, dependency-minimal message bus that lets multiple Claude Code sessions communicate in real-time over tmux, enabling one session to ask another a question and receive the answer without manual intervention.

AI & Machine LearningAI Agents
9
Vincentwei1021
Vincentwei1021 GitHub avatar

video-shotcraft

An AI agent skill that turns Claude Code or Codex into a motion-design studio for crafting cinematic product videos with Remotion, offering 106 shot recipe cards, 162 styles, 161 motion previews, and a production-ready template.

TypeScript
2,432
7-e1even
Featured
7-e1even GitHub avatar

learn-agent

A collection of notes on building coding agents, derived from the production development of the desktop agent Reina, with each mechanism simplified into a zero-dependency, single-file Node.js demo.

AI & Machine LearningLarge Language Models
218