# OECD MCP Server [Health: Active]

**Category:** 🧮 Data Science Tools  
**Repository:** https://github.com/isakskogstad/OECD-MCP-server  
**GitHub Stars:** 8  
**npm Downloads (last month):** 117  
**Views:** 1  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/oecd-mcp-server-2

## Description
OECD economic and statistical data via SDMX API. Access 5,000+ datasets across 17 categories.

## Tools
Capabilities this server exposes over MCP:

- **search_dataflows** — Search for OECD datasets (dataflows) by keyword. Returns matching datasets with their IDs, names, and descriptions.
- **list_dataflows** — List available OECD dataflows (datasets), optionally filtered by category. Use this to browse datasets by topic area.
- **get_data_structure** — Get the metadata and structure of a specific OECD dataset. Returns dimensions, attributes, and valid values for querying data.
- **query_data** — Query actual statistical data from an OECD dataset. ⚠️ IMPORTANT: Defaults to last 100 observations (max 1000) to protect context window. Use filters, time periods, or last_n_observations to control data size. Large datasets (e.g. SOCX_AGG) can have 70,000+ observations - always specify limits!
- **get_categories** — Get all available OECD data categories (17 categories covering all topics: Economy, Health, Education, Environment, etc.)
- **get_popular_datasets** — Get a curated list of commonly used OECD datasets across all categories.
- **search_indicators** — Search for specific economic or social indicators by keyword (e.g., "inflation", "unemployment", "GDP").
- **get_dataflow_url** — Generate an OECD Data Explorer URL for a dataset. Use this to provide users with a direct link to explore data visually in their browser.
- **list_categories_detailed** — Get all OECD data categories with example datasets for each category. Returns comprehensive information about all 17 categories.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "oecd-mcp-server": {
    "command": "npx",
    "args": ["-y","oecd-mcp"]
  }
}
```

## Documentation

## What OECD MCP Server MCP server does

OECD MCP Server MCP server gives MCP-compatible AI clients access to OECD economic and statistical data through the OECD SDMX API. The catalog covers more than 5,000 datasets across 17 categories, including quarterly national accounts, main economic indicators, health statistics, education data, and environmental indicators. The available data includes information from OECD member countries and partner economies.

The server supports both dataset discovery and data retrieval. An agent can search by topic, browse categories, look up indicators, inspect a dataset's dimensions and metadata, then submit a statistical query. It can also produce a link to the corresponding OECD Data Explorer view for sharing or follow-up analysis.

## How it works

The server communicates with an MCP client over stdio. A client starts the published npm package with `npx`, or launches the compiled Node.js entry point when running from source. Requests are handled through the OECD SDMX API rather than a local database, so the server acts as an MCP interface to OECD's catalog and statistical endpoints.

In addition to tools, the implementation exposes MCP resources for categories, popular dataflows, country codes, SDMX filter guidance, a glossary, AI-assistant instructions, and API information. It also includes prompt templates for trend analysis, country comparisons, latest statistics, dataset exploration, question-driven dataset discovery, filter construction, and Nordic comparisons.

## Setup and configuration

The simplest installation uses the published `oecd-mcp` package with `npx`. Add an MCP server entry that runs `npx -y oecd-mcp`, then restart the client. Claude Code can register the same command with `claude mcp add oecd npx -y oecd-mcp`, while OpenAI Codex uses a stdio entry in `~/.codex/config.toml`.

A source installation requires cloning the repository, running `npm install`, and building with `npm run build`. The client can then launch `node` with the absolute path to `dist/index.js`. No environment variables or API credentials are documented for this setup.

## Tools and capabilities

The nine documented tools are:

- `search_dataflows` searches datasets by keyword.
- `list_dataflows` filters datasets by category.
- `get_categories` returns the 17 available categories.
- `get_popular_datasets` returns commonly used datasets.
- `search_indicators` finds specific indicators.
- `get_data_structure` returns dataset metadata and structure.
- `query_data` retrieves statistical data.
- `get_dataflow_url` creates an OECD Data Explorer link.
- `list_categories_detailed` returns expanded category information.

Together, these tools support workflows that move from a broad research question to a selected dataflow, an understood filter structure, and a statistical result.

## Limitations and notes

The server depends on the OECD SDMX API, so data availability follows the datasets and coverage provided by OECD. The documented transport is stdio, and client setup examples are provided for Claude Desktop, Claude Code, and OpenAI Codex.

_Full upstream README: https://allmcps.com/mcp/oecd-mcp-server-2/readme_