🚀 Apprise-MCP

Give Your AI Agent Hands to Ping Your Phone, Discord & Slack

Never stare at your terminal waiting for Claude or Cursor again.
Let your AI text your phone the second it finishes a long task, encounters an error, or spots a market opportunity.

Features • Quick Start • Client Setup • Supported Channels • License

📱 Live Demo (Claude Desktop & Cursor Push Alert)

What you prompt your AI:
"Analyze this 50-page financial report, summarize the top 3 risks, and ping my Telegram when you're done."

What happens:
You close your laptop, grab a coffee, and your phone buzzes instantly with the full summary!

💡 Why Do You Need This?

Related MCP server: AgentPing

⚡ Quick Start (30 Seconds)

1. Install Dependencies

pip install apprise fastmcp pydantic

2. Verify Your First Alert

Run a quick test using your Telegram, Discord, or Slack webhook:

python -c "import apprise; ap = apprise.Apprise(); ap.add('tgram://YOUR_BOT_TOKEN/YOUR_CHAT_ID'); ap.notify(title='🎉 Hello!', body='Apprise-MCP is live!')"

🔌 Connect to Your AI Client

1. Claude Desktop / Cowork

Open your claude_desktop_config.json:

• Windows: %APPDATA%\Claude\claude_desktop_config.json

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

Add this block under mcpServers:

{
  "mcpServers": {
    "apprise-mcp": {
      "command": "python",
      "args": [
        "path/to/apprise-mcp/server.py"
      ]
    }
  }
}

2. Cursor IDE

1. Open Cursor Settings (Ctrl + Shift + J or Cmd + Shift + J) -> Features -> MCP.

2. Click + Add New MCP Server:

   • Name: apprise-mcp

   • Type: command

   • Command: python path/to/apprise-mcp/server.py

3. Hit Save. When the status turns green, Cursor's Composer can ping your phone anytime!

3. VS Code (Cline / Roo Code / Continue)

Add the same JSON snippet to your cline_mcp_settings.json.

4. OpenClaw & Agent Skill Framework

apprise-mcp includes a native skill definition package under skills/openclaw/SKILL.md.
Autonomous agents can install it to trigger notifications automatically on task completion or error escalation.

💻 Universal CLI for Terminal & Shell Scripts

apprise-mcp includes a standalone, zero-dependency CLI tool (cli.py) for developers, bash scripts, and headless agents:

# 1. Send single-channel alert
python cli.py notify -u "tgram://BOT_TOKEN/CHAT_ID" -t "Build Finished" -b "Deployment succeeded."

# 2. Multi-channel parallel broadcast (Telegram + Discord simultaneously)
python cli.py notify \
  -u "tgram://BOT_TOKEN/CHAT_ID" \
  -u "discord://WEBHOOK_ID/WEBHOOK_TOKEN" \
  -t "🚨 Critical Production Alert" \
  -b "Database memory above 95%."

# 3. JSON Output (optimized for AI agents and automated pipelines)
python cli.py notify -u "tgram://..." -t "Alert" -b "Message" --json

# 4. View popular channel syntax cheatsheet
python cli.py channels

🛠️ Popular Supported Channels

apprise-mcp inherits the full power of caronc/apprise, supporting over 80+ notification services and 160+ URL schemas.

👉 View Complete List of All 161 Supported Services & URL Schemas

Pass any combination of URLs, and your agent will broadcast to all of them simultaneously:

🧠 How the AI Uses It

Once connected, your AI Agent gains access to the send_notification tool.

┌──────────────────────────────────────────────┐
│                  AI Agent                    │
│   (Claude, Cursor, DeepSeek, VS Code, Auto)  │
└──────────────────────┬───────────────────────┘
                       │ Calls: send_notification(payload)
                       ▼
        ┌──────────────────────────────┐
        │       Apprise-MCP Server     │
        │    (FastMCP + Apprise Engine)│
        └──────────────┬───────────────┘
                       │ Dispatches simultaneously to:
       ┌───────────────┼───────────────┬───────────────┐
       ▼               ▼               ▼               ▼
  [ Telegram ]    [ Discord ]      [ Slack ]       [ Email ]
   (Your Phone)  (Dev Channel)   (Team Channel)  (Client Inbox)

Example Multi-Channel Agent Call:

{
  "service_urls": [
    "tgram://BOT_TOKEN/CHAT_ID",
    "discord://WEBHOOK_ID/WEBHOOK_TOKEN"
  ],
  "title": "🚨 Deploy Successful!",
  "body": "Version v1.2.0 has been deployed to production. All 42 tests passed."
}

🤝 Acknowledgements & Ecosystem

• Built on top of the battle-tested caronc/apprise library created by Chris Caron.

• Protocol powered by Model Context Protocol (MCP) and FastMCP.

📄 License

This project is licensed under the MIT License.