io.github.gjeltep/app-store-connect-mcp
Enables management of App Store Connect resources including app reviews, TestFlight crashes, analytics reports, and Xcode Cloud workflows through natural language or AI agents.
README
App Store Connect MCP Server
<!-- mcp-name: io.github.gjeltep/app-store-connect-mcp -->
Talk to App Store Connect about your app. Modular tools, async I/O, and OpenAPI‑driven typing so your agent stays accurate as Apple evolves.
Why this is different
- Spec‑aware: Fields and enums are derived from Apple’s OpenAPI spec at runtime, reducing drift and surprise breakage.
- Fast by default: Async
httpxclient, server‑side filtering, and smart pagination to keep payloads lean. - Smart filtering: Server‑side + client‑side filtering with chainable filter engine for complex queries.
- Modular domains: Clean separation of tool schemas and handlers; add new domains without touching the core; optimized LLM discovery/ usage
- MCP‑native: Stdio transport, capability declarations, and tool wiring align with the official SDK python‑sdk README.
Installation
From PyPI
pip install app-store-connect-mcp
From Source
git clone https://github.com/gjeltep/app-store-connect-mcp.git
cd app-store-connect-mcp
uv pip install -e ".[dev]"
Configuration
Set these environment variables:
# Required
export APP_STORE_KEY_ID="YOUR_KEY_ID"
export APP_STORE_ISSUER_ID="YOUR_ISSUER_ID"
export APP_STORE_PRIVATE_KEY_PATH="/path/to/AuthKey_XXXXX.p8"
# Optional
export APP_STORE_APP_ID="YOUR_APP_ID"
export APP_STORE_KEY_TYPE="team" # or "individual"
Usage
Run with your environment variables set (recommended):
app-store-connect-mcp
Or use a .env file during development:
# Copy and configure .env file
cp .env.example .env
# Edit .env with your credentials
# Run with --env-file flag
app-store-connect-mcp-dev --env-file .env
# Validate configuration without starting server
app-store-connect-mcp-dev --env-file .env --validate-only
Use with any MCP‑compatible client; the server announces tools and handles calls over stdio.
Generate or update API models
Models are generated from Apple's official OpenAPI spec (fetched automatically from Apple's developer site).
You can override the source with APP_STORE_CONNECT_OPENAPI_URL to point to a local JSON file.
uv pip install -e .[dev]
python scripts/generate_models.py
Development
For development setup, testing, and contribution guidelines, see CONTRIBUTING.md.
Tools
Tools use a resource-first naming convention (resource.verb) with category tags for discoverability.
App Tools
- reviews.list: List customer reviews with filters (
rating,territory,appStoreVersion). - reviews.search: Advanced search with rating ranges, territory matching, date windows, and content search.
- reviews.get: Get detailed review information.
TestFlight Tools
- crashes.list: List crash submissions from beta testers with filters (
device_model,os_version,app_platform,device_platform,build_id,tester_id). - crashes.search: Advanced search with:
- Server‑side filters (
appPlatform,deviceModel,osVersion) - Post‑filters: OS ranges (min/max), device model substrings (e.g., "iPhone 15"), and date windows (
created_since_days,created_after,created_before).
- Server‑side filters (
- crashes.get_by_id: Get detailed information about a specific crash submission.
- crashes.get_log: Retrieve the raw crash log text for a specific submission.
Analytics Tools
Requests:
- analytics_report_requests.list: List analytics report requests for an app with filters (
access_type). - report_requests.create: Create new analytics report requests for specific metrics and timeframes.
- report_requests.get: Get detailed information about a specific analytics report request.
Reports:
- report_requests.list_reports: List available reports within a request with filters (
name,category). - reports.get: Get specific analytics report information.
- reports.list_instances: List report instances with filters (
granularity,processing_date). - report_instances.get: Get detailed information about a specific report instance.
Segments:
- report_instances.list_segments: List data segments for a report instance.
- report_segments.get: Get segment download information (checksum, URL, size).
- report_instances.download_data: Download analytics report data to a TSV file.
Xcode Cloud Tools
Products & Workflows:
- products.list: List all Xcode Cloud products with filters (
product_type). - products.get: Get detailed information about a specific product.
- workflows.list: List workflows for a product with filters (
is_enabled). - workflows.get: Get detailed workflow information (note: create/update/delete operations not supported for safety).
Builds:
- builds.list: List builds for a product or workflow with filters (
execution_progress,completion_status,is_pull_request_build). - builds.get: Get detailed build information including status, duration, and issue counts.
- builds.start: Start a new build for a workflow with optional branch/tag or pull request specification.
Build Artifacts & Results:
- artifacts.list: List downloadable artifacts for a build.
- issues.list: List issues (errors, warnings) for builds or workflows.
- test_results.list: List test results including status, duration, and failure messages.
SCM Integration:
- scm_providers.list: List configured SCM providers (GitHub, GitLab, Bitbucket).
- repositories.list: List Git repositories for an SCM provider.
- pull_requests.list: List pull requests for a repository.
- git_references.list: List branches and tags for a repository.
Architecture
graph TD
%% Entry Point
subgraph "Entry Point"
SERVER[server.py<br/>MCP stdio server]
end
%% Domain Implementations
subgraph "Domain Handlers"
TESTFLIGHT[testflight/handlers.py<br/>TestFlight crash<br/>management]
APP[app/handlers.py<br/>App Store review<br/>management]
ANALYTICS[analytics/handlers.py<br/>Analytics report<br/>management]
XCODE_CLOUD[xcode_cloud/handlers.py<br/>Xcode Cloud CI/CD<br/>management]
end
%% Core Architecture
subgraph "Core Components"
subgraph "Core Framework"
PROTOCOLS[protocols.py<br/>Abstract interfaces]
BASE_HANDLER[base_handler.py<br/>Abstract base class]
QUERY_BUILDER[query_builder.py<br/>Fluent query construction]
FILTERS[filters.py<br/>Chainable filter engine]
RESPONSE_HANDLER[response_handler.py<br/>API response processing]
CONTAINER[container.py<br/>Dependency injection]
ERRORS[errors.py<br/>Structured error handling]
end
subgraph "API Layer"
ASC_CLIENT[app_store_connect.py<br/>App Store Connect client<br/>with JWT auth]
HTTP_CLIENT[http_client.py<br/>Base HTTP client]
MODELS[app_store_connect_models.py<br/>Auto-generated Pydantic<br/>models from OpenAPI spec]
end
subgraph "Data & Tools"
OPENAPI[app_store_connect_api_openapi.json<br/>Apple's OpenAPI spec]
GENERATE_SCRIPT[scripts/generate_models.py<br/>Model generation script]
end
end
%% Vertical Stack Relationships
SERVER -->|"routes to"| ANALYTICS
ANALYTICS -->|"uses"| ASC_CLIENT
%% Styling
classDef entry fill:#f3e5f5,stroke:#4a148c,stroke-width:2px
classDef core fill:#fff3e0,stroke:#e65100,stroke-width:2px
classDef domain fill:#fce4ec,stroke:#880e4f,stroke-width:2px
class SERVER entry
class PROTOCOLS,BASE_HANDLER,QUERY_BUILDER,FILTERS,RESPONSE_HANDLER,CONTAINER,ERRORS,ASC_CLIENT,HTTP_CLIENT,MODELS,OPENAPI,GENERATE_SCRIPT core
class TESTFLIGHT,APP,ANALYTICS,XCODE_CLOUD domain
Credits
Built on the official Model Context Protocol Python SDK — see the docs and examples in the python‑sdk README.
See CONTRIBUTING.md if you'd like to help.
推荐服务器
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。
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。
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。
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。
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。
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。
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。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
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。