World Bank Data360 MCP Server
# World Bank Data360 MCP Server
[](https://smithery.ai/server/@llnOrmll/world-bank-data-mcp)
MCP server that exposes World Bank Data360 OPEN API to Claude Desktop.
## Features
- 🔍 Search 1000+ economic and social indicators
- 📊 Access data for 200+ countries
- 📅 Historical data spanning 60+ years
- 🌍 Filter by country, year, demographics
## Installation
```bash
cd /Users/llnormll/WorkSpace/world-bank-mcp
# Install dependencies
uv sync
```
## Test
```bash
uv run world_bank_mcp/server.py
```
## Claude Desktop Configuration
Edit `~/Library/Application Support/Claude/claude_desktop_config.json`:
```json
{
"mcpServers": {
"world-bank-data": {
"command": "uv",
"args": [
"--directory",
"/path/to/world-bank-mcp",
"run",
"src/world_bank_mcp/server.py"
]
}
}
}
```
Restart Claude Desktop.
## Usage
Ask Claude:
- "Get world poverty data for 2024"
- "Show me GDP per capita for Japan in 2020"
- "Compare CO2 emissions for USA, China, and India"
Claude will automatically:
1. Search for the right dataset
2. Check available years
3. Retrieve the data with proper filters
4. Format and present results
## Available Tools
### 1. `search_datasets`
Search for datasets by keywords.
**Tip**: Use optimized queries like "gross domestic product total" instead of "GDP data"
### 2. `get_temporal_coverage`
Get available years for a specific dataset.
### 3. `retrieve_data`
Retrieve actual data with filters (year, countries, demographics).
## Data Sources
- **WB_WDI**: World Development Indicators
- **WB_HNP**: Health, Nutrition & Population Statistics
- **WB_GDF**: Global Development Finance
- **WB_IDS**: International Debt Statistics
## Common Country Codes
USA, CHN, JPN, DEU, GBR, FRA, IND, BRA, RUS, CAN, KOR, AUS, MEX, IDN, TUR, SAU, ARG, ZAF, ITA, ESP
## License
MITTDQS
Scored across 5 tools
Each tool has a clearly distinct purpose with no overlap: search_datasets_tool and search_local_indicators serve different search functions (API vs. offline), get_temporal_coverage_tool handles year availability, retrieve_data_tool fetches actual data, and list_popular_indicators provides curated discovery. The descriptions explicitly differentiate their roles, preventing agent misselection.
The naming is mixed, with some tools using verb_noun (search_datasets_tool, retrieve_data_tool) and others using noun_verb (get_temporal_coverage_tool, list_popular_indicators). While all names are readable and descriptive, the inconsistency in verb placement and suffix usage ('_tool' on some but not others) reduces predictability. The pattern is not chaotic but lacks uniformity.
With 5 tools, the count is well-scoped for a World Bank data server, covering essential workflows: discovery (search_local_indicators, list_popular_indicators), search (search_datasets_tool), validation (get_temporal_coverage_tool), and retrieval (retrieve_data_tool). Each tool earns its place without bloat, supporting a clear data access pipeline.
The tool set covers the core data retrieval workflow comprehensively: search, temporal validation, and data fetching, with additional discovery aids. Minor gaps exist, such as no explicit tools for filtering or aggregating data beyond basic parameters, but agents can work around these using the provided tools. The surface supports end-to-end data access without dead ends.