OpenViSTAI · Agents · Blockchain · Crypto
Menu
Multi-Agent Systems

Agent Message Queue

Agent Message Queue is a Go-based CLI and self-hosted developer tool that provides a Maildir-style file queue for communication between coding agents through a shared directory. It stores agent inboxes and outboxes locally, supports threaded replies and receipts, and can issue optional wake notifications. The queue uses plain files rather than a messaging server or database.

Independent OpenViST project profile, based on linked sources.

Agent Message Queue — OpenViST project profile in Multi-Agent Systems.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use AMQ when coding agents need a lightweight local channel for review requests, replies, or sequential handoffs and a shared directory is suitable for coordination. The documented workflow starts both agents before exchanging messages; earlier messages remain available for later draining.

Who it is for

  • Developers coordinating coding agents on a shared machine.
  • Teams using command-line agents such as Claude Code and Codex CLI.
  • Builders who want local, file-based inter-process communication between agent sessions.

Documented uses

  • Send a coding task or review request from one agent to another.
  • Read and drain an agent inbox while recording message receipts.
  • Reply within an existing message thread.
  • Create isolated sessions and route messages between agent mailboxes.
  • Connect selected running sessions to the optional Buzz remote companion.
Before you choose

Requirements & limits

The repository is public and licensed under MIT. The documented setup uses a CLI on macOS or Linux; the repository also describes WSL and native Windows paths with platform-specific distinctions.

  • A shared local directory is required for the core queue.
  • The core design does not use a centralized messaging server or database.
  • AMQ is a coordination layer, not a distributed broker.
  • Wake notifications prompt inbox checks but do not confirm that a request was read or completed.
  • The optional bridge and remote components are separate companions with their own documented workflows.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Set up an AMQ coding-agent session

Install AMQ and create a project session with agent mailboxes.

Before you start

  • macOS or Linux for the illustrated terminal workflow, or WSL on Windows with the Linux binary.
  • The relevant agent CLIs installed and available on PATH.
  • A project repository where the agents will work.

Steps

  1. Install the AMQ binary using the documented Homebrew or installer method, then confirm it with amq --version.

  2. Install the agent instructions with npx skills add avivsinai/agent-message-queue -g -y and restart the agents.

  3. Run amq setup in the project repository, select the agents and launcher, and confirm the preview.

  4. Run amq launch and accept the launch plan; with the commands backend, paste the printed complete command for each agent into a separate terminal.

Expected outcome

The project has a configured session and mailboxes, and the selected agents are started through the chosen launcher or terminal commands.

Guide limitations

  • The quick-start walkthrough does not start agents during its initial demo.
  • Messages sent before a recipient wake starts remain unread and do not trigger that new wake.
  • AMQ handles messages; task execution, worktrees, permissions, and merges remain under the user’s or orchestration tools’ control.
COMPARE OPTIONS

Compare options for AI coding assistance

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
Agent Message QueueSend a coding task or review request from one agent to another.CLI, Self-hostedThe repository is public and licensed under MIT. The documented setup uses a CLI on macOS or Linux; the repository also describes WSL and native Windows paths with platform-specific distinctions.
Claude Code Agent FarmRun multiple coding agents on codebase bug-fixing tasks.CLIThe project is available as a public GitHub repository and is used through a CLI.
Clawteam OpenclawCreate a team around a coding goal.CLIThe project is open source under the MIT license. It is installed from its GitHub repository using Python packaging.
Evidence and transparency

Sources & review

Source record details
Source availability
Active
Sources checked
4 October 2026
Original website title
avivsinai/agent-message-queue · Official homepage title · retained in its source language