ntfy-notify
Lets you send push notifications to your phone via ntfy, enabling alerts from scheduled tasks or direct messages.
README
ntfy-notify
A tiny Model Context Protocol (MCP) server that lets Claude / Cowork push notifications to your phone via ntfy. Point a scheduled task at it and the results land on your lock screen instead of waiting in a log somewhere.
Why I built this: I run a scheduled daily job-search inbox sweep, and I wanted the summary on my phone the moment it finishes instead of having to go check on it. ntfy is a dead-simple pub/sub-over-HTTP service, so the whole thing is one HTTP POST behind an MCP tool.
What it does
Exposes two MCP tools over stdio:
| Tool | Purpose |
|---|---|
send_notification(message, title="", priority="default", tags="", click_url="") |
Send any push notification to your configured ntfy topic. |
send_job_alert(summary) |
Convenience wrapper: title "Job sweep", high priority, briefcase tag. |
priority accepts min / low / default / high / urgent (or "1".."5").
tags is a comma-separated list of emoji shortcodes,
e.g. tada,computer. click_url opens when you tap the notification.
How ntfy works (30-second version)
ntfy is pub/sub over HTTP. You pick a topic — any string — and subscribe to it in the phone app. Anyone who knows the topic name can publish to it with a single HTTP POST, and it shows up on every subscribed device. There's no account or auth for public topics, which means the topic name is the only secret, so make it long and random.
How it works
There's no daemon and no always-on process. A stdio MCP server is launched by the client (Claude/Cowork) as a child process when a session starts, speaks JSON-RPC over stdin/stdout, stays idle until a tool is called, and is shut down when the session ends.
Claude / Cowork session
│ (launches as subprocess, JSON-RPC over stdio)
▼
server.py ──► @mcp.tool() functions
│ send_notification / send_job_alert
│ builds one HTTP POST:
│ body = message text
│ headers = Title / Priority / Tags / Click / Authorization
▼
https://ntfy.sh/<NTFY_TOPIC>
│ (pub/sub fan-out)
▼
ntfy app on your phone
Key design points:
- The topic is read from the environment, never hardcoded (
_config()). Since the topic name is the only secret in ntfy, this is what makes the repo safe to be public — and it fails loudly with a clear error ifNTFY_TOPICis unset. @mcp.tool()turns a plain function into a tool. The function name, type hints, and docstring become the schema and description that Claude reads to decide when and how to call it.- ntfy has no JSON payload. The message is the raw request body; everything else (title, priority, tags, click URL, auth token) rides along as HTTP headers.
- Every call returns an
OK:/FAILED:status string so the model knows whether the push actually went out. Network errors and non-2xx responses are caught and reported rather than raised.
Development
Run the test suite (pure logic + header construction, no network calls):
pip install -e ".[dev]"
pytest
1. Install the ntfy app and pick a topic
- Install ntfy on your phone:
- iOS: https://apps.apple.com/us/app/ntfy/id1625396347
- Android: Play Store or F-Droid
- Generate a long, random topic name so nobody else can guess it:
openssl rand -hex 16 - In the app, tap + and subscribe to that exact string.
2. Install the server
Requires Python 3.10+.
git clone https://github.com/Gardner-Programs/ntfy-notify.git
cd ntfy-notify
python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
3. Configure environment variables
Copy the example and fill it in:
cp .env.example .env
| Variable | Required | Default | Notes |
|---|---|---|---|
NTFY_TOPIC |
yes | — | Your long, random topic. Never hardcoded. |
NTFY_BASE_URL |
no | https://ntfy.sh |
Set to your own host if self-hosting. |
NTFY_TOKEN |
no | — | Bearer token for protected / self-hosted topics. |
The server reads these from the process environment. The MCP config below sets
them directly, so a .env file is only needed for the curl test / manual runs.
4. Register the server with Cowork / Claude
Add an entry to your MCP client config. For Claude Desktop
(claude_desktop_config.json) or Cowork, point it at the cloned repo:
{
"mcpServers": {
"ntfy-notify": {
"command": "python",
"args": ["/absolute/path/to/ntfy-notify/server.py"],
"env": {
"NTFY_TOPIC": "your-long-random-topic",
"NTFY_BASE_URL": "https://ntfy.sh",
"NTFY_TOKEN": ""
}
}
}
}
For Claude Code, the equivalent one-liner:
claude mcp add ntfy-notify \
--env NTFY_TOPIC=your-long-random-topic \
-- python /absolute/path/to/ntfy-notify/server.py
If you installed into a virtualenv, use that interpreter's absolute path (e.g.
/absolute/path/to/ntfy-notify/.venv/bin/python) as thecommand.
Restart the client and the send_notification / send_job_alert tools will appear.
5. Verify with curl
You don't need this server to test ntfy itself — confirm your topic works first:
curl \
-H "Title: Hello from ntfy" \
-H "Priority: high" \
-H "Tags: tada" \
-d "It works!" \
https://ntfy.sh/your-long-random-topic
With a token (protected / self-hosted):
curl \
-H "Authorization: Bearer tk_yourtoken" \
-d "It works!" \
https://ntfy.example.com/your-long-random-topic
You should get a push on your phone within a second or two.
Integration: daily job-inbox-sweep
Once the server is registered, update the scheduled daily-job-inbox-sweep task so its final step calls the tool with the run's summary, e.g.:
At the end of the run, call
send_job_alertwith a one-paragraph summary of what was found (new postings, replies, anything needing action).
send_job_alert already sets a sensible title and high priority, so a single call
with the summary text is all the task needs.
License
MIT — see LICENSE.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.