An MCP server that lets AI/LLM agents push and fetch rich ntfy notifications (supports self-hosted servers, token auth, URL/markdown auto-detection, etc.). Handy CLI & Docker images.
https://github.com/gitmotion/ntfy-me-mcpYour AI assistant just spent 20 minutes generating that complex codebase, but you missed the completion message because you were in another tab. Sound familiar?
ntfy-me-mcp bridges the gap between AI agents and real-world notifications. Your Claude or other MCP-compatible assistant can now send rich push notifications directly to your phone or desktop when tasks complete, errors occur, or milestones are reached.
You start a complex task with your AI assistant - maybe code generation, data analysis, or document processing. Then you context-switch to other work. When you check back, you've missed the completion, error messages, or important updates. You end up either:
This MCP server gives your AI agents two powerful capabilities:
Real-time push notifications - Your assistant can send notifications to any device with the ntfy app installed. No complex infrastructure, no API keys to manage (for basic use), just instant notifications.
Smart notification features - The server automatically detects URLs in messages and makes them clickable, enables markdown formatting when present, and supports priority levels, tags, and emoji.
Here's what a typical workflow looks like:
You: "Generate a React component for user authentication, then notify me when it's done"
[20 minutes later, notification pops up on your phone]
š± "Code Generation Complete ā
"
"Your React auth component is ready with TypeScript types and tests"
[View Code] [View Tests]
Add this to your MCP configuration and you're done:
{
"ntfy-me-mcp": {
"command": "npx",
"args": ["ntfy-me-mcp"],
"env": {
"NTFY_TOPIC": "your-unique-topic"
}
}
}
Install the ntfy app on your phone, subscribe to that topic, and your AI can start sending notifications immediately.
Code Generation: Get notified when large codebases are complete, with direct links to files or GitHub PRs.
Data Processing: Know instantly when data analysis, file processing, or API migrations finish.
Research Tasks: Receive summaries when document analysis, web research, or competitive analysis completes.
Error Monitoring: Get immediate alerts when your AI encounters issues during long-running tasks.
Build & Deploy: Notifications when CI/CD processes complete, with direct links to deployment previews.
The intelligent URL detection means your AI can include GitHub links, deployment URLs, or documentation links in messages - they automatically become clickable buttons in the notification.
Message Retrieval: Your AI can also fetch and filter past notifications, making it easy to review what happened while you were away.
Rich Formatting: Automatic markdown detection means code snippets, lists, and formatted text work seamlessly.
Private Servers: Full support for self-hosted ntfy instances with token authentication.
Priority Management: Critical errors get high-priority notifications that break through do-not-disturb settings.
Most notification solutions require complex setups, paid services, or custom infrastructure. ntfy-me-mcp works with the free ntfy.sh service or your own server, requires zero configuration beyond topic selection, and gives your AI agents native notification capabilities without writing a single line of notification code.
The automatic URL detection and markdown formatting mean your AI doesn't need special training or complex prompts - it just includes URLs or formatted text naturally, and the notifications work intelligently.
Works with any MCP-compatible AI assistant. Your Claude, Custom GPT, or other agent can start sending notifications with natural language:
The server handles the technical details while your AI focuses on the task at hand.
Ready to stop missing important updates from your AI workflows? The setup takes less time than reading this article.