Skip to main content

irm-kmi-mcp

License: MIT Python 3.10+ PyPI version CI

An MCP server exposing official Belgian weather data from the Royal Meteorological Institute of Belgium (IRM/KMI): observed conditions, daily and hourly forecasts, a radar-based rain nowcast, pollen levels and weather warnings, for MCP clients such as Claude, Hermes or Openclaw.

"What's the weather in Namur?"            →  current_conditions("Namur")
"Forecast for Oostende, 5 days, in Dutch" →  daily_forecast("Oostende", days=5, language="nl")
"Will it rain soon in Brussels?"          →  rain_forecast("Bruxelles")
"How bad is the pollen today?"            →  pollen()
"Any warnings in Belgium right now?"      →  warnings(country="BE")

irm-kmi-mcp on Glama

Features

  • Observed conditions: temperature, condition, UV index, sunrise/sunset from the nearest IRM weather station.
  • Daily forecast (1–8 days): min/max temperature, condition, wind, precipitation, sunrise/sunset, plus the official IRM text bulletin in fr/nl/en/de (can be omitted to save tokens).
  • Hourly forecast (1–49 h): temperature, precipitation, pressure, wind and gusts — every row carries an absolute ISO-8601 timestamp (Brussels time).
  • Rain nowcast: radar-based precipitation per 10-minute frame for roughly the next 5 hours; answers "will it rain soon?" far more cheaply than the hourly forecast.
  • Pollen levels: official daily pollen levels (grasses, birch, mugwort, …) when in season.
  • Official weather warnings: yellow/orange/red alerts, either for a single municipality (with level and validity period) or country-wide (filterable to BE, NL or LU).
  • Locations by name or coordinates: "Namur", "Bruxelles", "Oostende", … resolved through the IRM's own city search, or given directly as WGS84 latitude/longitude.
  • Token-lean output: unknown fields are omitted rather than sent as nulls.
  • Caching & resilience: weather data is cached for 10 minutes, city lookups for 24 hours; transient network errors are retried automatically.

Tools

Every location-taking tool accepts either commune (municipality name) or latitude + longitude (WGS84).

Tool Parameters (besides location) Returns
current_conditions — temperature, condition (ww code + canonical English label), UV index, day/night, sunrise/sunset
daily_forecast days (1–8, default 3), language (fr/nl/en/de), include_text (default true) per day: temps, condition (ww code + label), wind, precipitation, sunrise/sunset, official text
hourly_forecast hours (1–49, default 24) per hour: absolute time, temperature, condition (ww code + label), precipitation, pressure, wind
rain_forecast language unit, official IRM hint, precipitation per 10-min frame (only frames expecting rain)
pollen — per-species pollen level (none, active, low, moderate, high, very high), national
warnings language, country (BE/NL/LU, country-wide only) alert type (name + canonical slug), text; level + validity for a municipality, per-region level/validity otherwise

Data source & disclaimer

⚠️ Unofficial. This project is not affiliated with the IRM/KMI. It works by calling the same backend service that powers the official IRM mobile app. That service was never meant for public use: it is undocumented and it can break or disappear without notice. Please use it for personal, low-frequency purposes only.

The API surface was reverse-engineered and documented by Jules Dejaeghere in jdejaegh/irm-kmi-api and jdejaegh/irm-kmi-ha. Many thanks for the groundwork. This project is not affiliated with, sponsored or endorsed by the IRM/KMI.

Installation

From PyPI:

pip install irm-kmi-mcp

…or run it zero-install on every invocation with uvx (recommended for MCP clients — no venv to manage):

uvx irm-kmi-mcp

See Development below for installing from source.

Usage

Run the server (stdio transport):

irm-kmi-mcp
# or
python -m irm_kmi_mcp

Claude Desktop

Zero-install with uvx (recommended):

{
  "mcpServers": {
    "irm-kmi-mcp": {
      "command": "uvx",
      "args": ["irm-kmi-mcp"]
    }
  }
}

Or, after pip install irm-kmi-mcp:

{
  "mcpServers": {
    "irm-kmi-mcp": {
      "command": "irm-kmi-mcp"
    }
  }
}

To set a default language (see Configuration), add an "env" key to either form, e.g. "env": { "IRM_LANG": "fr" }.

Hermes Agent

After pip install irm-kmi-mcp (so the irm-kmi-mcp command is on PATH):

# config.yaml
mcp_servers:
  irm-kmi-mcp:
    command: "irm-kmi-mcp"
    enabled: true
    # env:
    #   IRM_LANG: "fr"  # optional: default language for official texts

No API keys or configuration required. The daily request key is derived automatically.

Configuration

  • IRM_LANG: default language for official texts (bulletins, warning texts, day names, wind directions): fr, nl, en or de. Unset or invalid values fall back to English. Read once at server startup; set it before launching the server.

Development

Install from source in an editable dev environment (requires uv):

git clone https://github.com/kthys/irm-kmi-mcp.git
cd irm-kmi-mcp
uv venv
uv pip install -e ".[dev]"
uv run --no-project pytest
uv run --no-project ruff check .

Tests run against recorded API responses (tests/fixtures/) with a mocked HTTP transport, so no network access is required. Re-record fixtures after upstream API changes.

Release files for irm-kmi-mcp 1.1.2

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for irm-kmi-mcp 1.1.2
File Size Uploaded
irm_kmi_mcp-1.1.2.tar.gz 34.1 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for irm-kmi-mcp 1.1.2
File Interpreter ABI Platform
irm_kmi_mcp-1.1.2-py3-none-any.whl Python 3 none any Details

Total release size: 59.5 kB

Release files / irm_kmi_mcp-1.1.2.tar.gz

Download URL irm_kmi_mcp-1.1.2.tar.gz
Size 34.1 kB
Tags Source
SHA-256 checksum
How to use checksums
e890f5b275e825734e4f9e7c56013355893cffd296d77bba6cb65395d387c2cb
BLAKE2b-256 checksum
How to use checksums
d433a89101a935ad5b1aa65e2a3e340fc071d64b6d8233a283443444aaa9e5ce
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 28, 2026.

Transparency log

Release files / irm_kmi_mcp-1.1.2-py3-none-any.whl

Download URL irm_kmi_mcp-1.1.2-py3-none-any.whl
Size 25.5 kB
Tags Python 3
SHA-256 checksum
How to use checksums
3756ff53dbf5d4e73667f90c751197b306dc494be4eae43d21e31fd0806c4616
BLAKE2b-256 checksum
How to use checksums
a260d72fd8977c543541ab55e9593e419ffed987ae1b248e5a2e870d0d209105
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 28, 2026.

Transparency log

Release history Release notifications | RSS feed

This release

1.1.2 This release

2 release files

1.1.1

2 release files

1.1.0

2 release files

1.0.1

2 release files

1.0.0

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page