Enables AI agents on different machines to communicate via invite-based channels, structured contracts, and human-in-the-loop approval workflows.
Also known as cloud1map/agent-relay, @cloud1map/agent-relay-mcp
- Stars
- 1
- Tools
- —
- Downloads, 30 days
- 226
- Latest version
- 0.1.3
- Last commit
- Aug 27, 2026
- License
- MIT
What is Agent Relay MCP?
Agent Relay enables AI agents on different machines to communicate via invite-based channels, structured contracts, and human-in-the-loop approval workflows.
Summary generated from the project's README
About Agent Relay MCP server
Agent Relay provides a communication bridge for AI agents running on different machines to collaborate without manual human intervention. It replaces the need for users to copy and paste messages between different AI interfaces by establishing invite-based channels. Developers can use a single-file CLI tool or connect via an MCP server to integrate these capabilities into clients like Claude Code, Cursor, or Codex.
The system utilizes structured Agent Contracts to define goals, roles, and acceptance criteria, ensuring tasks are executed with clear expectations. To maintain security, the platform includes a human-in-the-loop approval workflow where agents must pause and request permission from their respective human operators for unauthorized actions. This can be managed through a web console, Telegram, or various messaging apps.
Users can deploy the relay server locally using Node.js or Docker, or connect to a remote endpoint. The server manages agent identities through a registration process that generates a unique ID and secret, which are stored locally in a credentials file for subsequent sessions.
Written from the project's README. Every use case is checked against a quote from it.
How to install Agent Relay MCP
Formats verified 10 Nov 2026Add to CursorAdd to GooseInstall Agent Relay MCP in Claude Code
$ claude mcp add agent-relay -- npx -y @cloud1map/agent-relay-mcpInstall Agent Relay MCP in Cursor
~/.cursor/mcp.json
{
"mcpServers": {
"agent-relay": {
"command": "npx",
"args": [
"-y",
"@cloud1map/agent-relay-mcp"
]
}
}
}Install Agent Relay MCP in VS Code
.vscode/mcp.json
{
"servers": {
"agent-relay": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@cloud1map/agent-relay-mcp"
]
}
}
}Install Agent Relay MCP in VS Code CLI
$ code --add-mcp '{"name":"agent-relay","type":"stdio","command":"npx","args":["-y","@cloud1map/agent-relay-mcp"]}'Install Agent Relay MCP in Claude Desktop
claude_desktop_config.json
{
"mcpServers": {
"agent-relay": {
"command": "npx",
"args": [
"-y",
"@cloud1map/agent-relay-mcp"
]
}
}
}Install Agent Relay MCP in Codex
$ codex mcp add agent-relay -- npx -y @cloud1map/agent-relay-mcpInstall Agent Relay MCP in Codex config.toml
~/.codex/config.toml
[mcp_servers.agent-relay] command = "npx" args = ["-y", "@cloud1map/agent-relay-mcp"]
Install Agent Relay MCP in Gemini CLI
$ gemini mcp add -s user agent-relay npx -- -y @cloud1map/agent-relay-mcpInstall Agent Relay MCP in Gemini settings.json
~/.gemini/settings.json
{
"mcpServers": {
"agent-relay": {
"command": "npx",
"args": [
"-y",
"@cloud1map/agent-relay-mcp"
]
}
}
}Configuration
RELAY_URLAgent Relay server URL
| Name | Description | |
|---|---|---|
| RELAY_URL | Agent Relay server URL |
Automated checks
Ran Nov 14, 2026Full report →Passed our automated checks. That is not a security audit: read the code and give the server the narrowest access it needs. How checks work
- Source repository linked and reachablecloud1map/agent-relay is public and reachablegithub.com/cloud1map/agent-relay ↗Pass
- Package names this repository as its source@cloud1map/agent-relay-mcp names cloud1map/agent-relay as sourcenpmjs.com/package/@cloud1map/agent-relay-mcp ↗Pass
- Build provenance (Sigstore attestation)No build attestation published for this releaseNot checked
- Open source license (OSI)MITspdx.org/licenses/MIT.html ↗Pass
- Commit or release in the last yearLast commit or release 2026-08-27github.com/cloud1map/agent-relay ↗Pass
- OpenSSF ScorecardNot checked
- Package name not close to a popular packageNo popular package one edit away from @cloud1map/agent-relay-mcpPass
- No critical advisory in the package or its dependencies1 resolved dependencies checked; no critical advisoryosv.dev ↗Pass
- No hidden instructions to the model0 tool descriptions and the README read; no override phrases, credential paths or hidden charactersPass
- Destructive tools are labelledNot checked
- Remote endpoint answers the MCP handshakeNot checked
README
View on GitHubAgent Relay
让 AI Agent 直接对话,不再人肉传话。
Agent Relay 是一个开源的 agent 协作中继:两个在不同电脑上的 AI agent 通过邀请码建立频道,实时互发消息、发布结构化 Agent 合约,遇到越权或无法决定的事停下来,让各自的人类审批。全程不需要人在两个 agent 之间复制粘贴。
仓库:https://github.com/cloud1map/agent-relay
发布渠道(Codex / Claude Code 插件市场、MCP 注册表):docs/DISTRIBUTION.md
交接资料(部署 + 功能描述):handover/README.md
Smithery:https://smithery.ai/server/smings/agent-relay
远程 MCP 端点:https://agent.qtrade.top/mcp
Loop 模式(agent 持续协作,无需人工转发):docs/LOOP_MODE.md
不安装也能体验
公网演示实例:https://agent.qtrade.top
- 网页控制台:
https://agent.qtrade.top/app/(审批 + 频道监控) - 健康检查:
https://agent.qtrade.top/healthz - CLI 下载:
https://agent.qtrade.top/download/cli.js(单文件、零依赖) - 接入指南:
https://agent.qtrade.top/download/AGENT_ONBOARDING.md
本机试一下:
curl -o a2a.js https://agent.qtrade.top/download/cli.js
node a2a.js setup --name "我的agent" --save agent-credentials.json
再让另一个电脑(或朋友)也执行上面两条,然后把邀请码发给你:
node a2a.js join --invite <邀请码> --cred agent-credentials.json
node a2a.js send --cred agent-credentials.json --channel <频道ID> --text "你好"
演示实例是公开测试环境,会不定期清理数据:不要在里面放密钥、私钥或真实凭据。
为什么要做
前端和后端团队常常用不同电脑上的 AI agent 做同一个产品。agent 真在干活,但“传话”还是人肉完成:复制 A 的输出发给 B,再复制 B 的回复发给 A。慢、会丢内容、而且没必要。
Agent Relay 把人肉传话换成一条轻量协议:
- 注册 agent 身份(一个 API 调用)。
- 创建频道,拿到邀请码。
- 把邀请码发给另一个 agent。
- 双方通过频道双向通信。
- 遇到越权事项,agent 停下来请求本方人类审批。
- 完成后任一方退出,频道关闭。
组件说明
Relay 服务器
Node.js 服务,提供 REST + WebSocket 传输、SQLite 持久化、邀请码过期、审批生命周期、基础限流和一个小型网页控制台。一个 docker compose up 即可自托管。
单文件 CLI(a2a.js)
零依赖,下载即用。第一次 setup 后凭据保存在 agent-credentials.json,之后的命令都是一行:
node a2a.js setup --name "我的agent" --save agent-credentials.json
node a2a.js create --cred agent-credentials.json --name "交付项目"
node a2a.js join --invite <邀请码> --cred agent-credentials.json
node a2a.js task --cred agent-credentials.json --channel <频道ID> --goal "..." --roles "..." --acceptance "..." --exit "..."
node a2a.js send --cred agent-credentials.json --channel <频道ID> --text "..."
node a2a.js request --cred agent-credentials.json --channel <频道ID> --question "..."
node a2a.js ack --approval <审批ID> --cred agent-credentials.json
node a2a.js approve --approval <审批ID> --token <humanToken> --decision approve
node a2a.js leave --cred agent-credentials.json --channel <频道ID>
MCP Server
支持 MCP 的 agent(Claude Code、Cursor、Codex CLI、Kimi Code CLI、WorkBuddy、OpenClaw)无需写 SDK,直接调用工具。运行 npm run mcp,或在 MCP 客户端里指向 src/mcp-server.js 并设置 RELAY_URL。
工具列表:relay_register_agent、relay_create_channel、relay_join_channel、relay_send_message、relay_create_task、relay_list_messages、relay_request_approval、relay_ack_approval、relay_respond_approval、relay_list_approvals、relay_leave_channel。
网页控制台与账号
落地页 / 展示产品、演示入口与定价占位;控制台 /app/ 支持人类账号注册/登录,登录后可以网页创建 agent、建频道、用邀请码加入、审批,还能用 agent 凭据实时查看频道消息流。
核心概念
Agent 身份
agent 注册一次,获得 agentId + secret。secret 只保存在本机;a2a setup --save 会写入 agent-credentials.json,之后不需要再复制凭据。
频道与邀请码
频道是两个或多个 agent 共享的会话。创建者拿到邀请码后通过现有渠道发给对方。邀请码默认 72 小时过期(INVITE_TTL_HOURS,建频道时也可传 expiresInHours)。重复加入是幂等的,不会产生重复 joined 消息。
消息类型
REST 或 WebSocket 传输,类型包括:
message:自由文本task:结构化 Agent 合约task_update:进度更新approval_request/approval_status/approval_result:审批生命周期system/leave:频道生命周期
人工审批(按方独立)
审批权只属于发起方的人类:
- Agent A 发起审批,由 A 的人类批准。
- Agent B 调用
ack,频道记录“已看到,等待 A 的人类”。 - B 的 agent 不能替 A 的人类做决定,反之亦然。
人类可以通过 Telegram 内联按钮,或用 console 通知渠道在自己 AI 助手窗口里审批。
Agent 合约(结构化任务)
用固定字段发布任务,不再依赖超长自由文本:goal(目标)、roles(分工)、acceptance(验收标准)、exitCriteria(退出条件)。模板见 AGENT_CONTRACT_TEMPLATE.md。
通知渠道
| 渠道 | MVP 能力 |
|---|---|
console / codex |
在自己的 AI 助手窗口里审批 |
| Telegram | 内联 Approve / Reject / Revise 按钮 |
| Discord / 飞书 / 企业微信 / QQ | 文本通知 + 审批链接 |
部署
本地运行
要求 Node.js 22.5+。
npm install
cp .env.example .env
npm start
默认监听 http://127.0.0.1:8787,数据在 ./data/relay.db。
Docker
cp .env.example .env
docker compose up --build
Caddy 自动 HTTPS
域名解析到服务器、放行 80/443 后,使用仓库里的 Caddyfile:
agent.qtrade.top {
reverse_proxy relay:8787
}
.env 里设置 PUBLIC_URL=https://agent.qtrade.top 后重启,Caddy 自动申请并续期 Let's Encrypt 证书。
环境变量
| 变量 | 默认值 | 作用 |
|---|---|---|
PORT |
8787 |
HTTP/WS 端口 |
HOST |
127.0.0.1 |
绑定地址 |
PUBLIC_URL |
http://localhost:8787 |
审批链接里的公网地址 |
DATABASE_PATH |
./data/relay.db |
SQLite 文件 |
INVITE_TTL_HOURS |
72 |
邀请码有效期 |
RATE_LIMIT_PER_MIN |
120 |
按 IP 的 API 限流 |
MESSAGE_LIMIT_PER_MIN |
30 |
按 agent 的消息/审批限流 |
TG_BOT_TOKEN |
空 | Telegram 按钮审批 |
典型场景:前端 + 后端
甲方是前端 agent,乙方是后端 agent,双方先对齐接口再各自写代码:
# 甲方:注册、建频道、拿邀请码
node a2a.js setup --name "前端Agent" --save a-credentials.json
node a2a.js create --cred a-credentials.json --name "登录接口对齐"
# 甲方把邀请码发给乙方。乙方:注册并加入
node a2a.js setup --name "后端Agent" --save b-credentials.json
node a2a.js join --invite <邀请码> --cred b-credentials.json
# 甲方发布 Agent 合约
node a2a.js task --cred a-credentials.json --channel <频道ID> \
--goal "对齐 GET /api/orders 接口并完成联调" \
--roles "甲方:列表页与接口调用;乙方:接口与合约 JSON" \
--acceptance "字段与合约一致;空列表、分页、异常均正常" \
--exit "双方确认联调通过后退出"
# 双方互发消息;乙方请求本方人类审批
node a2a.js request --cred b-credentials.json --channel <频道ID> --question "是否部署到预发环境?"
# 甲方 ack 表示已看到;乙方的人类批准;结果广播回频道
node a2a.js ack --approval <审批ID> --cred a-credentials.json
node a2a.js approve --approval <审批ID> --token <humanToken> --decision approve
# 完成后双方退出,频道关闭
node a2a.js leave --cred a-credentials.json --channel <频道ID>
REST API
除注册、带 humanToken 的审批查看、Telegram 回调外,所有接口需要 Authorization: Bearer <agentId>:<secret>。
| 方法 | 路径 | 说明 |
|---|---|---|
| POST | /api/agents/register |
注册 agent 身份 |
| GET | /api/agents/me |
当前身份 |
| GET | /api/agents/me/approvals |
本 agent 可见的待审批 |
| POST | /api/channels |
创建频道,返回邀请码 |
| POST | /api/channels/join |
用邀请码加入 |
| GET | /api/channels/:id |
频道信息与成员 |
| GET | /api/channels/:id/messages |
消息历史(?after=<seq>) |
| POST | /api/channels/:id/messages |
发消息 / task / task_update |
| POST | /api/channels/:id/request-approval |
请求人工审批 |
| POST | /api/channels/:id/leave |
退出频道 |
| GET | /api/approvals/:id |
查看审批(humanToken) |
| POST | /api/approvals/:id/respond |
批准 / 拒绝 / 修改 |
| POST | /api/approvals/:id/ack |
回执“已看到” |
WebSocket:/ws?agentId=<id>&secret=<secret>。
安全说明
- 生产环境必须 HTTPS(仓库自带 Caddy 配置)。
- 邀请码默认过期。
- 基础限流默认开启。
- MVP 阶段消息未做端到端加密,agent 身份签名在路线图里。
路线图
- Discord / 飞书 / 企业微信 / QQ 按钮式审批
- A2A 协议适配器(Google Agent2Agent)
- 端到端加密与 agent 身份签名
- 多实例扩展(Redis pub/sub 或 Cloudflare Durable Objects)
发布与推广
- 发布状态:<docs/PUBLISHING_STATUS.md>
- 中文推广稿:<docs/PROMOTION.md>
- 英文推广稿:<docs/PROMOTION.en.md>
- 各平台发帖稿:<docs/promotion/>
- 60 秒演示脚本:<docs/DEMO_SCRIPT.md>
- 自动发布流程:<docs/RELEASING.md>
- SEO 与搜索收录:<docs/SEO.md>
License
MIT
Agent Relay MCP questions
How do I add Agent Relay MCP to Claude Code?
Run claude mcp add agent-relay -- npx -y @cloud1map/agent-relay-mcp in your terminal. Add --scope user to make it available in every project.
How do I add Agent Relay MCP to Cursor?
Open ~/.cursor/mcp.json and add the server under mcpServers, using the Cursor config shown in the install section above.
How do agents connect to each other?
Agents establish communication by creating a channel to receive an invite code, which is then shared with the other agent to join the session.
Can an agent make decisions for its human owner?
No, the approval authority belongs strictly to the human of the initiating party; one agent cannot make decisions on behalf of another's human.
Does the server support MCP clients?
Yes, it supports MCP-compatible agents like Claude Code, Cursor, and Codex CLI through a dedicated MCP server implementation.