mobile-release-mcp

mobile-release-mcp

MCP server for managing mobile app releases on App Store Connect and Google Play. It exposes 101 tools for iOS and Android release lifecycle and reacts to EAS/GitHub webhooks automatically.

Category
访问服务器

README

<div align="center">

<img src="docs/assets/banner.jpg" alt="mobile-release-mcp — Ship iOS & Android from your AI agent" width="100%" />

<br />

101 MCP tools · App Store Connect · Google Play · EAS & GitHub webhooks

<br />

npm version CI License: MIT Node.js MCP

<br />

npx -y mobile-release-mcp · Quick start · Tools · Credentials · Changelog

</div>


Install

# Run directly (recommended for MCP clients)
npx -y mobile-release-mcp

# Or install globally
npm install -g mobile-release-mcp
mobile-release-mcp

No deployment required for local use with Cursor or Claude Desktop.

Quick start

1. Get credentials

You need App Store Connect and/or Google Play API credentials. EAS webhooks do not replace these — see docs/CREDENTIALS.md.

2. Connect to Cursor or Claude Desktop

{
  "mcpServers": {
    "mobile-release": {
      "command": "npx",
      "args": ["-y", "mobile-release-mcp"],
      "env": {
        "APPLE_KEY_ID": "ABC1234DEF",
        "APPLE_ISSUER_ID": "00000000-0000-0000-0000-000000000000",
        "APPLE_PRIVATE_KEY_PATH": "/path/to/AuthKey.p8",
        "MCP_TOOLSET": "release"
      }
    }
  }
}

3. Ask your agent

"List my App Store apps and the latest build for each one."

Alternative: run from source

git clone https://github.com/Jeronimo0228/mobile-release-mcp.git
cd mobile-release-mcp
npm install
cp .env.example .env
# Edit .env with your credentials
npm run build
npm start

For local development from source, point MCP clients to node /path/to/dist/index.js instead of npx.

Why use this

Benefit What you get
Agent-native releases Agents call typed tools instead of fragile shell scripts around fastlane or manual store UI clicks
Cross-platform One MCP server for App Store Connect and Google Play
Webhook-driven CI EAS/GitHub events trigger agent workflows via list_pending_webhooks
Production-ready defaults Retries on rate limits, pagination, persistent webhooks, startup validation
Safe by default Destructive tools require confirm: true; toolsets limit agent surface area

Features

  • 101 MCP tools — iOS, Android, and shared release operations
  • Webhook listener — EAS (Build + Submit) and GitHub Actions (workflow_run)
  • Signature verification — HMAC-SHA1 (EAS), HMAC-SHA256 (GitHub); secrets required in production
  • Persistent webhooks — events saved to disk (WEBHOOK_STORAGE_PATH), survive restarts
  • Unified release flowtrigger_full_release for both platforms
  • Stdio + HTTP transports — local MCP clients or remote /mcp deployment
  • Configurable toolsetsall, release, or readonly via MCP_TOOLSET
  • Structured errors — consistent JSON errors with retry hints for agents

Documentation

Doc Contents
docs/CREDENTIALS.md Store credentials vs webhook secrets, EAS project mapping
docs/TOOLSETS.md Limiting tools for agents (MCP_TOOLSET)
docs/TOOLS.md Full tool catalog with categories
docs/ARCHITECTURE.md Transports, reliability, project layout
docs/SECURITY.md Supply chain, HTTP auth, deployment hardening
SECURITY.md Vulnerability reporting

Configuration

Copy .env.example to .env when running from source, or set env vars in your MCP client config.

Minimum for iOS:

APPLE_KEY_ID=ABC1234DEF
APPLE_ISSUER_ID=00000000-0000-0000-0000-000000000000
APPLE_PRIVATE_KEY_PATH=/path/to/AuthKey.p8
MCP_TOOLSET=release

All environment variables

Variable Default Description
APPLE_KEY_ID App Store Connect API key ID
APPLE_ISSUER_ID App Store Connect issuer ID
APPLE_PRIVATE_KEY_PATH Path to .p8 key file
APPLE_PRIVATE_KEY_BASE64 Alternative: base64-encoded key
GOOGLE_SERVICE_ACCOUNT_KEY_PATH Path to service account JSON
GOOGLE_SERVICE_ACCOUNT_JSON Alternative: inline JSON
MCP_TRANSPORT stdio stdio or http
MCP_HTTP_API_KEY Required in HTTP mode (≥32 chars). Bearer token for /mcp
MCP_ALLOWED_ORIGINS Comma-separated CORS origins (empty = disabled)
MCP_PORT 3000 HTTP server port (when http)
MCP_TOOLSET all all, release, or readonly
WEBHOOK_PORT 3000 Webhook port in stdio mode
WEBHOOK_STORAGE_PATH .data/webhooks.json Persistent webhook storage
WEBHOOK_REQUIRE_SECRETS true Reject webhooks without secrets
EAS_WEBHOOK_SECRET EAS HMAC secret
GITHUB_WEBHOOK_SECRET GitHub HMAC secret
EAS_PROJECT_MAPPINGS JSON map EAS project → store IDs
LOG_LEVEL info debug, info, warn, error

Requirements

Important: EAS webhooks notify you about builds — they do not provide store API credentials. See docs/CREDENTIALS.md.

HTTP deployment (remote MCP)

MCP_TRANSPORT=http
MCP_PORT=3000
EAS_WEBHOOK_SECRET=your-secret
GITHUB_WEBHOOK_SECRET=your-secret

Endpoints on a single server:

Endpoint Method Purpose
/mcp GET, POST, DELETE MCP Streamable HTTP transport
/webhook/eas POST EAS build/submit webhooks
/webhook/github POST GitHub Actions webhooks
/health GET Health check

Connect MCP clients to http://your-host:3000/mcp.

Deploy with npx mobile-release-mcp as the start command and env vars from your host's secret manager (Railway, Fly.io, etc.).

Webhook setup

EAS

eas webhook:create --event BUILD --url https://your-server.com/webhook/eas --secret your-eas-secret
eas webhook:create --event SUBMIT --url https://your-server.com/webhook/eas --secret your-eas-secret

GitHub Actions

Settings → Webhooks → Payload URL: https://your-server.com/webhook/github, events: Workflow runs.

EAS project mapping

Map Expo project names to store identifiers:

EAS_PROJECT_MAPPINGS=[{"projectName":"my-app","iosAppId":"1234567890","androidPackageName":"com.example.app"}]

Webhook flow

EAS/GitHub → POST /webhook/* → verify signature → persist to disk
  → Agent: list_pending_webhooks (includes mappedTargets)
  → Agent: trigger_full_release (confirm: true)
  → Agent: mark_webhook_processed

Examples

Release on both platforms

{
  "platforms": ["ios", "android"],
  "confirm": true,
  "ios": {
    "appId": "1234567890",
    "buildId": "abc-build-id",
    "versionString": "2.1.0",
    "releaseNotes": [{ "locale": "en-US", "whatsNew": "Bug fixes" }]
  },
  "android": {
    "packageName": "com.example.app",
    "versionCodes": ["42"],
    "track": "production",
    "status": "completed"
  }
}

React to EAS build webhook

1. list_pending_webhooks → check mappedTargets.appleAppId
2. apple_list_builds → verify build is VALID
3. trigger_full_release with confirm: true
4. mark_webhook_processed

See docs/TOOLS.md for the complete tool reference.

Development

npm run dev        # Run with tsx
npm run typecheck  # TypeScript check
npm run build      # Production build
npm test           # Run tests

Publishing instructions: docs/PUBLISHING.md

Project structure

src/
├── index.ts              Entry point
├── server.ts             MCP server factory
├── http/app.ts           HTTP transport (MCP + webhooks)
├── providers/apple/      App Store Connect API
├── providers/google/     Google Play Developer API
├── tools/                MCP tool registrations
├── webhook/              Webhook routes, storage, parsers
└── utils/                Config, retry, errors, tool registry
docs/                     Detailed documentation
tests/                    Unit tests

License

MIT — see LICENSE.

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选