sg-property-mcp
# sg-property-mcp
**Singapore property prices, land use, and neighborhood amenities via MCP — no API keys needed.**
An [MCP](https://modelcontextprotocol.io/) server for Singapore property research. Works out of the box with zero configuration.
## What it does
| Tool | Description |
|------|-------------|
| **search_area** | Land use and zoning info near a Singapore address |
| **search_area_by_coords** | Same as above, using coordinates directly |
| **search_hdb_resale** | Recent HDB resale flat prices by town (2017 onwards) |
| **search_nearby_amenities** | Schools, MRT, parks, hawker centres, hospitals within a radius |
| **analyze_results** | AI-powered analysis of your last search |
| **export_csv** / **export_md** | Save results to CSV or Markdown files |
**8 tools. Zero API keys. Just works.**
## Quick start
### Desktop app
Add to your MCP client config:
```json
{
"mcpServers": {
"sg-property": {
"command": "npx",
"args": ["-y", "sg-property-mcp"]
}
}
}
```
### CLI (MCP-compatible)
```bash
# Example for CLI tools that support MCP:
npx sg-property-mcp
```
### Manual install
```bash
npm install -g sg-property-mcp
```
Then add to your MCP client config:
```json
{
"mcpServers": {
"sg-property": {
"command": "sg-property-mcp"
}
}
}
```
## Example conversations
> **"What's the land use around Bishan MRT?"**
> → `search_area` returns zoning types, plot ratios, planning areas
> **"Show me recent 4-room HDB prices in Tampines"**
> → `search_hdb_resale` returns transactions with prices, block, storey, lease info
> **"What amenities are within 500m of 1.3521, 103.8198?"**
> → `search_nearby_amenities` finds schools, MRT stations, parks, hawker centres with distances
> **"Analyze the price patterns in those results"**
> → `analyze_results` provides AI-generated insights on pricing trends
## Data sources
All data comes from free, public APIs. No registration or API keys required.
- Land use zoning — (c) Urban Redevelopment Authority
- HDB resale prices — Contains information from data.gov.sg accessed under the Singapore Open Data Licence
- Geocoding & amenities — Data (c) OpenStreetMap contributors
## Requirements
- Node.js 20.6 or later
- An MCP-compatible client (any MCP desktop app or CLI)
## License
MIT
TDQS
Scored across 8 tools
Tools are mostly distinct with clear purposes. Export_csv and export_md overlap but differ in format; search_area and search_area_by_coords serve same purpose with different inputs but descriptions clarify. Overall, minimal confusion.
All tool names follow a consistent verb_noun pattern in snake_case, e.g., search_area, export_csv, get_attributions. No deviations or mixed styles.
With 8 tools covering land use, HDB resale, amenities, analysis, exports, and attributions, the scope is well-focused for a property server. Not too many or too few.
Core workflows (search, export, analysis) are covered, but there are notable gaps: no direct property address lookup, no private property transaction data, and no filtering or sorting options. Agents might face dead ends.