Log inSubmit server
Unicef logo

Unicef

Unicef MCP server provides agent access to UNICEF datasets covering child health, education, protection, mortality, and related indicators.

InstallRepositoryWebsite
Full README Compare alternatives Embed widget Claim this listing

Install Unicef

Pick your client, then add it in one click or paste the config.

STDIONode.js
~/Library/Application Support/Claude/claude_desktop_config.json
claude_desktop_config.json
{
  "mcpServers": {
    "unicef": {
      "command": "npx",
      "args": [
        "-y",
        "@pipeworx/mcp-unicef"
      ]
    }
  }
}

Settings → Developer → Edit Config, merge this into mcpServers, then fully quit and reopen Claude Desktop.

Install not yet sandbox-tested— we haven't run this command through our automated sandbox yet. Not a red flag; we're working through the catalog.

Overview

The Unicef MCP server connects AI clients to UNICEF Data and exposes tools for finding datasets, inspecting their SDMX dimensions, and retrieving observations. Use `list_dataflows` to locate datasets, `dataflow_structure` to learn valid dimension codes, and `get_data` to query filtered series. It is suited to research and analysis involving child health, nutrition, education, mortality, immunization, WASH, and child-related SDGs. The server is available through a hosted MCP endpoint or as a local stdio package.

Use cases

  • Find UNICEF datasets by topic
  • Inspect valid country and indicator codes
  • Retrieve child mortality time series
  • Compare UNICEF observations across dimensions
  • Query child-related SDG data

Key features

  • Searches UNICEF dataflows by keyword
  • Exposes SDMX dimension structures
  • Retrieves filtered UNICEF observations
  • Supports wildcard dimension positions
  • Available as hosted MCP or local stdio

Tools3

What Unicef lets your agent do. Expand a tool for its parameters.

Self-reported~235 tokens of context

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked 3d ago

Recent health check succeeded.

9/9 checks healthy over the last 50d

Supply-chain signal

No high-severity advisories surfaced by our automated scan.

Critical 0High 0Medium 0Low 0

Scanned 2d ago via OSV.dev · @pipeworx/mcp-unicef (npm)

Adoption & maintenance

Factual signals from GitHub, npm, and our automated checks — not a rating.

npm downloads796Package downloads in the last 30 days.
Last commit29d agoMost recent push to the default branch.
Tools exposed3Callable tools this server registers over MCP.

How Unicef works

What the Unicef MCP server does

The Unicef MCP server exposes UNICEF Data through three focused MCP tools. It supports discovery of roughly 70 UNICEF datasets, inspection of each dataset’s structure, and retrieval of observations. The available subject areas include child health, nutrition, education, child protection, child mortality, child poverty, immunization, water, sanitation and hygiene, and child-related Sustainable Development Goals.

This makes the server useful when an agent needs source data rather than general discussion. For example, an agent can find a mortality dataset, identify its country and indicator codes, then request the relevant time series.

How it works

The tools follow a discovery-to-query sequence. First, call list_dataflows, optionally providing a keyword such as mortality or nutrition. Each result includes a dataflow ID, version, and English name.

Next, pass that ID to dataflow_structure. The response describes the dataset’s ordered dimensions and the valid codes within them, such as countries, indicators, sex, age, or wealth quintile. This order is required for the final query.

Finally, call get_data with the dataflow ID and a dot-separated SDMX key. Each key position corresponds to one dimension from dataflow_structure; an empty position acts as a wildcard. The response contains decoded series, dimension labels, and values for each period. Omitting the key or passing an empty value requests all series and may return a large result.

Setup and configuration

The Unicef MCP server can be connected through the hosted endpoint https://gateway.pipeworx.io/unicef/mcp. The README provides configuration examples for Claude Desktop, Cursor, and Windsurf. The same gateway also exposes a full endpoint at https://gateway.pipeworx.io/mcp.

For local use, run the package with npx -y @pipeworx/mcp-unicef and configure the MCP client to start that command. The local process communicates over standard input and output and exposes only the UNICEF pack’s three tools. No Pipeworx account or gateway round-trip is required for this mode.

The tools can also be called over HTTP. For example, the gateway exposes POST https://gateway.pipeworx.io/v1/tools/unicef_list_dataflows with a JSON body containing an optional query.

Tools and capabilities

  • list_dataflows: Search or browse UNICEF dataflows.
  • dataflow_structure: Inspect dimension order and valid codes for a selected dataflow.
  • get_data: Retrieve observations using SDMX dimension filters and wildcards.
  • Query topics including mortality, nutrition, education, immunization, WASH, and child-related SDGs.

Limitations and notes

The Unicef MCP server requires the dimension order and valid codes before constructing reliable get_data keys. Broad requests can produce large responses. Hosted pack connections also include shared Pipeworx meta-tools in addition to the three UNICEF tools; the exact tool list is reported by the connection’s initialize response. The local package does not include those shared meta-tools. The project is licensed under MIT.

Read the full README View source on GitHub

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about Unicef

Unicef is an MCP server that connects AI clients to UNICEF Data and exposes three main tools: list_dataflows for finding datasets, dataflow_structure for inspecting dimensions and valid codes, and get_data for retrieving filtered observations across child health, education, mortality, nutrition, and related topics.

Install it locally by configuring your MCP client to run `npx -y @pipeworx/mcp-unicef`. This starts the package as a stdio server and does not require a Pipeworx account. You can instead configure the hosted endpoint at `https://gateway.pipeworx.io/unicef/mcp`.

Yes, Unicef works with Cursor through the hosted MCP endpoint. Add `https://gateway.pipeworx.io/unicef/mcp` to Cursor’s MCP configuration, or run the local npm package through Cursor’s stdio configuration.

Call `dataflow_structure` before `get_data` to learn the dataset’s dimension order and valid codes. Then provide `get_data` with the dataflow ID and a dot-separated SDMX key, leaving positions empty when you want wildcard filters.

Related MCP servers

All in Data Science Tools All alternatives
Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewUnicef AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/unicef?style=directory)](https://allmcps.com/mcp/unicef)
HTML Embed
<a href="https://allmcps.com/mcp/unicef"><img src="https://allmcps.com/api/badge/unicef?style=directory" alt="Unicef on AllMCPs" /></a>