Self-hosted Node.js and React app for orchestrating teams of AI agents, with org charts, task tracking, budgets and governance.
AgentWorkforce/relay
Infrastructure that lets coding agents like Claude and Codex message each other, share integrations, and run multi-step workflows across machines.
- Stars
- 1,097
- Forks
- 97
- Open issues
- 337
- Last commit
- 3 hours ago
What is it?
Agent Relay is a toolkit for teams that run coding agents together. It adds a coordination layer on top of the agents and tools already in use, rather than replacing them. Agents can send messages through shared channels, threads, DMs, and files, and they can run on different machines while staying in the same workspace.
The project also exposes external tools such as GitHub, Linear, Notion, and Slack as a virtual filesystem, so agents can read and write them with ordinary commands like ls, cat, and grep. A Shared Sessions component captures coding agent sessions so teammates and their agents can search earlier work, decisions, and context through an MCP connection.
Flows let developers define multi-step workflows in TypeScript with deterministic checks, required steps, and human approval gates. The project offers a hosted cloud option, an install script and npm CLI, and self-hosting guides for each component.
Key features
- 01Agents from different tools and machines can message each other through shared channels, threads, and DMs.
- 02External services like GitHub, Linear, Notion, and Slack are accessible as a virtual filesystem.
- 03Captured coding sessions can be searched by teammates and their agents.
- 04Multi-step workflows are defined in TypeScript with deterministic checks and human gates.
- 05Components can be self-hosted or used through the hosted cloud.
Use cases
- Coordinating several coding agents working on the same codebase across different machines.
- Enforcing required review or CI steps in agent-driven pull request workflows.
- Letting a team search past agent sessions for earlier decisions and context.
- Letting agents read and update issues or pages in GitHub, Linear, or Notion via file commands.
Quick start
- 1
Run the official install script from the README
- 2
Or install the CLI with npm
npm install -g agent-relay
Things to know
- Full setup often relies on Agent Relay Cloud or a self-hosting process that the README points to but does not detail.
- Some integrations and features depend on Cloud login or separate projects.
In collections
Topics
FAQ
- What is relay?
- Infrastructure that lets coding agents like Claude and Codex message each other, share integrations, and run multi-step workflows across machines. Agent Relay is a toolkit for teams that run coding agents together. It adds a coordination layer on top of the agents and tools already in use, rather than replacing them. Agents can send messages through shared channels, threads, DMs, and files, and they can run on different machines while staying in the same workspace.
- Is relay free to use? What license does it have?
- relay is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
- Is relay still maintained?
- Actively maintained. The last commit to relay was on Nov 7, 2026.
- What can I use relay for?
- Coordinating several coding agents working on the same codebase across different machines. Enforcing required review or CI steps in agent-driven pull request workflows. Letting a team search past agent sessions for earlier decisions and context. Letting agents read and update issues or pages in GitHub, Linear, or Notion via file commands.
- How do I install relay?
- Run the official install script from the README Or install the CLI with npm: npm install -g agent-relay
- Where can I get relay?
- The source code and installation instructions are on GitHub: https://github.com/AgentWorkforce/relay
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
An open collaboration system where people and their AI agents share context, coordinate tasks, and work in parallel within shared Spaces.
A self-hosted dashboard to orchestrate Claude Code, Codex and other AI agents, with persistent memory, agent-to-agent messaging and multi-machine support.
A local-first desktop workbench that runs OpenCode, Codex and DeepSeek Harness agents on shared projects, tasks, plugins and Skills.
Open-source Python framework for orchestrating role-based autonomous AI agents (Crews) and event-driven workflows (Flows).
A CLI and workflow engine that runs coding agents like Claude Code, Codex and Copilot through configurable review, repair and test gates before work is accepted.