Model Context Protocol server

ICOS greenhouse-gas data, ready for Responsible AI

Find atmosphere, ocean and ecosystem data, size a selection, compute statistics on the server and fetch the data — with the citation, licence and DOI of the data sources in every answer, and rich contextual metadata on quality and provenance, fully traceable through data passports.

MCP is an open standard that lets AI assistants such as Claude or ChatGPT use outside tools and data.

Connect your assistant Explore the zarr datasets
https://mcp.icos-cp.eu/mcp · streamable HTTP
6zarr datasets
3domains: atmosphere, ecosystem, ocean
8tools and 3 guided prompts
CC BY 4.0licence, DOI in every answer

Connect in one line

One endpoint for any MCP client. The server is a thin layer over the ICOS zarr service: every number comes from there, and every data answer carries its RO-Crate data passport.

Claude Code
claude mcp add --transport http icos https://mcp.icos-cp.eu/mcp
Claude Desktop, VS Code and other clients
{"mcpServers": {"icos": {"type": "http", "url": "https://mcp.icos-cp.eu/mcp"}}}

Three domains of the carbon cycle

What is in the air, what land ecosystems exchange with it, and what the ocean takes up — measured with standardised methods across the ICOS station network.

Find, size, compute, fetch, cite

Eight tools that follow how a scientist works with data — and keep heavy computation on the server, next to the data.

1Discover

list_datasetsList the ICOS datasets: one per data store, with domain (atmosphere, ocean, ecosystem), time and space coverage, licence and DOI.
search_variablesFind variable ids in the ICOS catalogue by name, quantity or unit.
describe_variableOne variable in full: names, unit, which dataset, its coverage, how its quality flag is read, licence and DOI.

2Locate & size

find_stationsThe stations that carry a variable: id, name, position, sampling height, country, ecosystem -- near a point (sorted by distance), in a box, by name or by country.
estimateHow big a selection is -- rows, stations, time span -- without fetching it.

3Compute

aggregateStatistics over a WHOLE selection, computed on the server: count, mean, min, max, sum and standard deviation of one variable, overall or per station, year, month or day.

4Fetch & cite

get_dataObservations for one variable, filtered by box, time, station, cruise, height and quality control.
get_citationHow to cite a variable's data: the citation text, licence and DOI from its data passport, and -- for a selection (stations, a time window, a box) -- every data object it came from, in full.

Guided prompts

station_recordPlot or describe one station's time series.
compare_stationsCompare one quantity across several stations.
ocean_regionWhat surface-ocean CO2 observations exist in a region.

Behind every number, a station team

The values an assistant retrieves here are produced by people in the field and the lab, following common protocols and quality control across the ICOS network.

Two ICOS station staff in climbing harnesses next to a measurement mastTwo researchers checking an instrument enclosure at a coastal siteScientist in a laboratory

Every number carries its passport

Climate decisions need traceable evidence. Each data answer includes the citation, licence and DOI from its RO-Crate data passport — near-real-time data cite the exact data object they came from — so an assistant can pass them on.

SourceICOS zarr serviceQuality-controlled observations, one store per dataset.
This serverICOS MCPSelects, aggregates and attaches the data passport.
Every answerData + citation
citationCC BY 4.0DOI

ICOS data are licensed under CC BY 4.0. What the passport records: the data passport page.