OmniWire-MCP
Enables AI models to fetch and aggregate news from RSS, Atom, JSON, and HTML feeds with fault-tolerant circuit breaker protection.
README
OmniWire MCP Server ⚡️
OmniWire-MCP is a production-ready, fault-tolerant news aggregation server built on the Model Context Protocol (MCP). It bridges the gap between AI models and real-time information by providing a unified interface for consuming RSS, Atom, JSON, and HTML feeds.
Designed for resilience, it features a Sentinel Service with Circuit Breakers to handle failing sources gracefully and a Universal Parser that automatically adapts to different content formats.
✨ Features
- Universal Parsing: Automatically detects and parses RSS, Atom, JSON, and HTML content.
- Sentinel Architecture: Intelligent Circuit Breakers monitor source health, preventing cascading failures.
- Dynamic Configuration: Hot-reload sources via remote JSON config (
RSS_FEEDS) or fallback to local defaults. - Data Normalization: Transforms diverse feed formats into a standardized
NewsItemschema. - AI-Native: Exposes specialized Tools and Prompts optimized for LLM consumption.
- Zero-Config Deployment: Runs instantly via
npxor Docker.
🚀 Quick Start
Using npx (Global)
Prerequisite: The package must be published to NPM or installed globally.
npx omniwire-mcp
Local Testing (Without Publishing)
If you are developing locally and want to test the npx command:
- Link the package globally:
cd OmniWire-MCP npm link - Run with
npx:npx omniwire-mcp
Custom Configuration
You can configure the server in two ways:
-
Remote URL: Provide a URL to a JSON configuration file.
RSS_FEEDS="https://raw.githubusercontent.com/furkankoykiran/OmniWire-MCP/refs/heads/main/src/config/defaults/feeds.json" npx omniwire-mcp -
Direct JSON: Pass the configuration JSON directly as a string.
RSS_FEEDS='{"sources": [...]}' npx omniwire-mcp
📦 Installation & Publishing
Publish to NPM (Recommended)
To make npx omniwire-mcp available to everyone:
- Login to NPM:
npm login - Publish the package:
npm publish --access public
Option 1: Docker (Cloud/Production)
The server is Docker-ready for easy deployment.
# Build the image
docker build -t omniwire-mcp .
# Run with custom config
docker run -e RSS_FEEDS="https://raw.githubusercontent.com/furkankoykiran/OmniWire-MCP/refs/heads/main/src/config/defaults/feeds.json" omniwire-mcp
Option 2: Local Development
-
Clone the repository:
git clone https://github.com/furkankoykiran/OmniWire-MCP.git cd OmniWire-MCP -
Install dependencies and build:
npm install npm run build -
Run the server:
npm start
🔌 MCP Configuration
Add OmniWire to your MCP client configuration (e.g., claude_desktop_config.json or mcp_config.json).
Standard Configuration (npx)
{
"mcpServers": {
"omniwire": {
"command": "npx",
"args": [
"-y",
"@furkankoykiran/omniwire-mcp"
],
"env": {
"RSS_FEEDS": "https://raw.githubusercontent.com/furkankoykiran/OmniWire-MCP/refs/heads/main/src/config/defaults/feeds.json",
"LOG_LEVEL": "info"
}
}
}
}
Docker Configuration
{
"mcpServers": {
"omniwire": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e", "RSS_FEEDS",
"omniwire-mcp"
],
"env": {
"RSS_FEEDS": "https://raw.githubusercontent.com/furkankoykiran/OmniWire-MCP/refs/heads/main/src/config/defaults/feeds.json"
}
}
}
}
🛠 Capabilities
Tools
| Tool | Description | Arguments |
|---|---|---|
fetch-news |
Smat fetcher with Sentinel protection | filter (string), sourceId (string), limit (number) |
check-health |
Diagnostic report for sources | sourceId (optional) |
refresh-config |
Force reload of remote config | None |
reset-source |
Manually reset a circuit breaker | sourceId (string) |
Resources
| URI | Description |
|---|---|
news://all |
Aggregated feed from all healthy sources |
news://source/{id} |
Live feed from a specific source |
health://sources |
Real-time system health report |
config://current |
View active configuration |
Prompts
| Prompt | Application |
|---|---|
summarize-news |
"Give me a digest of topic X" (Uses fetch-news) |
analyze-sources |
"Diagnose my feed health" (Uses check-health) |
⚙️ Configuration Schema
The configuration JSON file should match the following schema:
{
"sources": [
{
"id": "tech-crunch",
"name": "TechCrunch",
"url": "https://techcrunch.com/feed/",
"type": "rss",
"enabled": true
}
],
"configPollIntervalMs": 60000,
"requestTimeoutMs": 10000,
"sentinel": {
"failureThreshold": 3,
"recoveryTimeoutMs": 60000
}
}
🤝 Contributing
We welcome contributions! Please see our Contributing Guide for details on how to set up your development environment and submit Pull Requests.
📄 License
This project is licensed under the MIT License.
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。