How do I use notification-mcp?

1. Click on "Deploy Server". 2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state. 3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@notification-mcp Let me know when the video export is done." That's it! The server will respond to your query, and you can continue using it as needed. Here is a step-by-step guide with screenshots.

notification-mcp

by elad12390

Overview Schema Source Related Servers Discussions Admin

TypeScript

Local

notification-mcp

An MCP server that lets AI assistants send spoken audio notifications using high-quality AI voices. When an AI finishes a long task, it can call the notify tool to speak a message aloud — so you don't have to stare at the screen waiting.

Powered by Kokoro TTS (fast, expressive, 26 voices), with optional F5-TTS (higher quality, voice cloning) and macOS say (instant, no setup).

Quick Start

Prerequisites

macOS (notifications play via afplay)
Node.js 18+

uv — Python package runner (for Kokoro/F5-TTS)

curl -LsSf https://astral.sh/uv/install.sh | sh

Install via npx

No installation needed — run directly with npx:

{
  "mcpServers": {
    "notification-mcp": {
      "command": "npx",
      "args": ["-y", "@elad12390/notification-mcp"]
    }
  }
}

Install from source

git clone https://github.com/elad12390/notification-mcp.git
cd notification-mcp
npm install
npm run build

Then configure your MCP client (see Configuration below).

Related MCP server: ntfy-me-mcp

Configuration

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "notification-mcp": {
      "command": "npx",
      "args": ["-y", "@elad12390/notification-mcp"]
    }
  }
}

OpenCode / Cursor / Other MCP Clients

{
  "mcpServers": {
    "notification-mcp": {
      "command": "npx",
      "args": ["-y", "@elad12390/notification-mcp"]
    }
  }
}

From source

{
  "mcpServers": {
    "notification-mcp": {
      "command": "node",
      "args": ["/path/to/notification-mcp/dist/index.js"]
    }
  }
}

The `notify` Tool

Once configured, the AI can call:

notify(message: "Hey! Build succeeded, all tests passed.")
notify(message: "Done. Your file has been exported.", title: "Export Complete")
notify(message: "Hmm... something looks off. Check the logs.", voice: "george")

Writing Expressive Messages

Kokoro reads punctuation expressively. Write like you'd want it spoken:

Punctuation	Effect	Example
`!`	Excited/energetic	`"Great job!"`
`...`	Long pause, thoughtful	`"Wait... what?"`
`,`	Brief natural pause	`"Well, I think so"`
`?`	Rising intonation	`"Really?"`
`.`	Sentence break	`"Done. Everything worked."`

Good examples:

"Hey! Your task is complete. Everything went smoothly."
"Hmm... I found something interesting. Check this out!"
"Done! The build succeeded, and all 47 tests passed."

Custom Pronunciation (IPA)

For words that need specific pronunciation, use [word](/IPA/] syntax:

"[Chaim](/χaim/) said hello"     → Hebrew name with guttural ח
"[Elad](/ɛlɑd/) is working"      → Pronounce as "eh-lahd"  
"The [route](/ruːt/) is clear"    → British-style pronunciation

Common IPA symbols: χ (Hebrew ח), ʃ (sh), ð (the), θ (think), ŋ (ng), ɹ (American r)

Voices

Kokoro Voices (default engine)

Use friendly aliases or full voice names:

Alias	Full Name	Accent
`heart`	`af_heart`	American Female
`bella`	`af_bella`	American Female
`nicole`	`af_nicole`	American Female
`sarah`	`af_sarah`	American Female
`adam`	`am_adam`	American Male
`michael`	`am_michael`	American Male
`emma`	`bf_emma`	British Female
`george`	`bm_george`	British Male

Additional full names: af_alloy, af_aoede, af_jessica, af_kore, af_nova, af_river, af_sky, am_echo, am_eric, am_fenrir, am_liam, am_onyx, am_puck, bf_alice, bf_isabella, bf_lily, bm_daniel, bm_fable, bm_lewis

TTS Engines

Set NOTIFICATION_METHOD to choose your engine:

Engine	Speed	Quality	Requires
`kokoro` (default)	~3–5s	Great, expressive	`uv`
`f5tts`	~10–15s	Highest, voice cloning	`uv` + FFmpeg
`say`	Instant	Robotic	Nothing (macOS built-in)

Environment Variables

Variable	Default	Description
`NOTIFICATION_METHOD`	`kokoro`	TTS engine: `kokoro`, `f5tts`, or `say`
`KOKORO_VOICE`	`af_heart`	Default Kokoro voice (alias or full name)
`KOKORO_SPEED`	`1.0`	Kokoro speech speed multiplier
`F5TTS_SPEED`	`1.0`	F5-TTS speech speed multiplier
`F5TTS_REF_AUDIO`	—	Path to reference audio file for voice cloning
`F5TTS_REF_TEXT`	—	Transcript of the reference audio
`SAY_VOICE`	`Samantha`	macOS `say` voice name
`LOG_LEVEL`	`info`	Log level: `debug`, `info`, `warn`, `error`, `silent`
`MCP_METRICS_ENABLED`	`true`	Enable/disable usage metrics

Example: Use a British male voice by default

{
  "mcpServers": {
    "notification-mcp": {
      "command": "npx",
      "args": ["-y", "@elad12390/notification-mcp"],
      "env": {
        "KOKORO_VOICE": "bm_george",
        "KOKORO_SPEED": "1.1"
      }
    }
  }
}

Example: Use F5-TTS with custom voice cloning

{
  "mcpServers": {
    "notification-mcp": {
      "command": "npx",
      "args": ["-y", "@elad12390/notification-mcp"],
      "env": {
        "NOTIFICATION_METHOD": "f5tts",
        "F5TTS_REF_AUDIO": "/path/to/your-voice.wav",
        "F5TTS_REF_TEXT": "Transcript of what is said in that audio file."
      }
    }
  }
}

Development

# Install dependencies
npm install

# Development with hot reload
npm run dev

# Build
npm run build

# Run tests
npm test

# Run with coverage
npm run test:coverage

License

MIT

Available Tools

2 tools

notifyA

Sends a notification to the user. Use this when you complete a task and want to alert the user.

IMPORTANT - Writing expressive messages:

Use punctuation naturally: "Hello!" sounds excited, "Hello..." sounds thoughtful
Use commas for pauses: "Well, I think so" has a natural pause
Use ellipsis for longer pauses: "Wait... what?"
Exclamation marks add energy: "Great job!" vs "Great job."
Question marks affect intonation: "Really?" sounds curious
Break long messages into sentences for natural rhythm
The model reads punctuation expressively, so write like you'd want it spoken!

Example good messages:

"Hey! Your task is complete. Everything went smoothly."
"Hmm... I found something interesting. Check this out!"
"Done! The build succeeded, and all tests passed."

CUSTOM PRONUNCIATION (IPA): For words that need specific pronunciation, use: [word](/IPA phonemes/) Common IPA symbols:

χ = Hebrew ח (chet), guttural "kh" sound
ʃ = "sh" sound
ð = "th" in "the"
θ = "th" in "think"
ŋ = "ng" sound
ɹ = American "r"

Examples:

"Chaim" - Hebrew name with guttural ח
"Elad" - Pronounce as "eh-lahd"
"Hello world!" - Custom pronunciation for "world"

TTS ENGINE OPTIONS (set via NOTIFICATION_METHOD env var):

"kokoro" (default): Fast (~3-5s), built-in voices, IPA pronunciation support
"f5tts": Higher quality (flow matching), slower (~10-15s), voice cloning capable
"say": macOS built-in, fastest but robotic

Inputs: message (required), reasoning (required), title, voice.

report_issueA

Report a bug, feature request, or other issue with the MCP server. Issues are saved locally for the server maintainer to review.

Inputs: title (required), reasoning (required), description (required), category, +5 more.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

Deploy Server

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

TDQS

A3.7/5.0

Scored 2026-09-13 13:55 across 2 tools

Disambiguation5/5

report_issue and notify serve clearly different purposes: one reports bugs/features about the MCP server, the other sends user-facing notifications. An agent can easily choose the correct tool based on task context.

Naming Consistency4/5

Both names use lowercase snake_case and start with a verb, but report_issue follows a verb_noun pattern while notify is a lone verb. This is a minor inconsistency, not confusing.

Tool Count3/5

With only two tools, the set is thin for a dedicated server, though both tools are distinct and useful. It falls into the borderline 1-2 tool range noted in the rubric.

Completeness4/5

The server covers the core notification action and provides a feedback path via report_issue. Minor gaps exist around notification management or delivery status, but these are not essential for basic use.

Maintenance

ActivityInactive

ResponsivenessNo recent community issues

Resources

Related MCP Connectors

VoiceLabsA
now.voicelabs
AI voice generation: text-to-speech and voice cloning from any MCP client.
Audio Processing Speech Processing Text-to-Speech
181
https://app.voicelabs.now/api/mcp
Danushkumar-V-mcp-discord
ai.smithery
An MCP server that integrates with Discord to provide AI-powered features.
Communication Social Media App Automation
2
https://server.smithery.ai/@Danushkumar-V/mcp-discord/mcp
Texttospeech McpC
io.github.varvararatta
MCP server for Text-to-Speech
Text-to-Speech Speech Processing Audio Processing
2
https://texttospeech-mcp.mcpize.run/mcp
mcp-aichat
io.github.AceDataCloud
MCP server for AI dialogue using various LLM models via AceDataCloud
AI & Machine Learning
2
https://aichat.mcp.acedata.cloud/mcp

Related MCP Servers

Sound Notification MCP
Code Execution Audio Processing
ks0318-p
A
license
Not graded
quality
D
maintenance
An MCP server that plays notification sounds when AI coding assistants like Windsurf or Cursor require user attention, such as when coding is complete or when user approval is needed.
Updated 2 years ago (2025-03-15 05:12 UTC)
17 npm
2
MIT
ntfy-me-mcp
Communication App Automation
gitmotion
A
license
A
quality
B
maintenance
A streamlined MCP server that enables AI assistants to send real-time notifications to your devices through the ntfy service, allowing you to receive alerts when tasks complete or important events occur.
Updated a month ago (2026-10-08 00:57 UTC)
2
244 npm
76
GPL 3.0
webhook-mcp
Communication Developer Tools
noobnooc
F
license
Not graded
quality
D
maintenance
An MCP server that sends webhook notifications to a configured URL, enabling AI assistants to notify users when long-running tasks complete.
Updated 2 years ago (2025-05-01 05:16 UTC)
29 npm
3
-
mcp-speak-when-done
Text-to-Speech Speech Processing
j3k0
A
license
A
quality
D
maintenance
An MCP server that makes AI agents speak a brief summary of every response out loud using TTS.
Updated 8 months ago (2026-03-22 09:39 UTC)
1
GPL 3.0

notification-mcp

Quick Start

Prerequisites

Install via npx

Install from source

Configuration

Claude Desktop

OpenCode / Cursor / Other MCP Clients

From source

The notify Tool

Writing Expressive Messages

Custom Pronunciation (IPA)

Voices

Kokoro Voices (default engine)

TTS Engines

Environment Variables

Example: Use a British male voice by default

Example: Use F5-TTS with custom voice cloning

Development

License

Available Tools

Tool Schema Changelog

TDQS

Maintenance

Resources

Related MCP Connectors

Related MCP Servers

Sound Notification MCP

ntfy-me-mcp

webhook-mcp

mcp-speak-when-done

The `notify` Tool