MCP Releases Server
An MCP server that provides GitHub release information through the Model Context Protocol. It features a tool to retrieve the latest stable release version of any specified GitHub repository.
README
MCP Releases Server
An MCP (Model Context Protocol) server that provides GitHub release information and PyPI package versions.
Features
- GitHub Releases: Query the latest stable release from any GitHub repository
- PyPI Integration: Look up the latest production version of Python packages
- Smart Fallback: Automatically falls back to PyPI when GitHub releases aren't available
- Repository Aliases: Configure short aliases for commonly used repositories
Tools
get_latest_release- Returns the latest stable release version of a GitHub repository (with PyPI fallback)get_pypi_version- Directly query PyPI for the latest stable version of any Python package
Configuration
Environment Variables
Create an optional .env file in the project root to configure the GitHub token:
export GITHUB_TOKEN="github_pat_TOKEN"
The GitHub token is optional but recommended to avoid API rate limits. The server will work without it, but with lower rate limits.
Repository Mapping
Configure repository aliases and PyPI packages in repo_mapping.json:
{
"fastapi": {
"repo": "tiangolo/fastapi",
"aliases": ["FastAPI"],
"pypi_package": "fastapi"
},
"nomad": {
"repo": "hashicorp/nomad",
"aliases": ["Nomad"]
}
}
Fields:
repo: GitHub repository inowner/nameformat (required)aliases: List of alternative names for this repository (optional)pypi_package: PyPI package name for fallback lookups (optional)
Running with Docker
docker-compose up --build
The server will be available at http://localhost:8000/mcp.
Running Locally
uv sync
uv run uvicorn server:app --host 0.0.0.0 --port 8000
VSCode Integration
Add the following to your VSCode settings (.vscode/mcp.json or user settings):
{
"mcp": {
"servers": {
"releases-server": {
"url": "http://localhost:8000/mcp"
}
}
}
}
Or add to ~/.vscode/mcp.json:
{
"servers": {
"releases-server": {
"url": "http://localhost:8000/mcp"
}
}
}
After adding the configuration, restart VSCode or reload the window. The get_latest_release and get_pypi_version tools will be available to Copilot.
PyPI Integration
How It Works
The server includes two ways to query PyPI:
- Direct PyPI Lookup: Use
get_pypi_version()to directly query PyPI for any package - Automatic Fallback: When
get_latest_release()can't find a GitHub release, it automatically checks PyPI if apypi_packageis configured inrepo_mapping.json
PyPI Features
- Filters out pre-release and dev versions
- Returns only stable production versions
- Includes package metadata (summary, homepage, URLs)
- Handles non-existent packages gracefully
Usage Examples
Direct PyPI lookup:
from server import get_pypi_version
# Query any package directly
result = get_pypi_version("django")
print(result)
# Output:
# Latest stable version for django (from PyPI):
# Version: 6.0.1
# Summary: A high-level Python web framework...
# PyPI URL: https://pypi.org/project/django/6.0.1/
With automatic fallback:
from server import get_latest_release
# Configure in repo_mapping.json:
# {
# "my-package": {
# "repo": "owner/repo",
# "pypi_package": "my-package"
# }
# }
# If GitHub fails, automatically falls back to PyPI
result = get_latest_release("my-package")
Test PyPI integration:
uv run python3 -c "
from server import get_latest_pypi_version
result = get_latest_pypi_version('requests')
print(f\"Version: {result['version']}\")
"
Running Integration Tests
To run the integration suite using pytest (dependencies are handled ephemerally by uv). The tests will automatically load the GITHUB_TOKEN from your .env file if present:
uv run pytest test_integration.py -v
Or run specific test categories:
# Test GitHub functionality
uv run pytest test_integration.py -k "github" -v
# Test PyPI functionality
uv run pytest test_integration.py -k "pypi" -v
# Test with verbose output
uv run pytest test_integration.py -v -s
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。