DVF Paris
An MCP app that displays interactive Paris real estate price statistics per arrondissement inside Claude, with map, comparison, and address search.
README
DVF Paris — MCP App
An interactive MCP App that displays Paris real estate prices per arrondissement, directly inside Claude.
Built with the MCP Apps SDK (@modelcontextprotocol/ext-apps).

What it does
Ask Claude about Paris real estate prices and get an interactive widget with:
- Fullscreen mode (v0.8): split layout with map (60%) and info panel (40%), interactive search
- Interactive map (Leaflet + OpenStreetMap) highlighting the arrondissement
- Choropleth sections (v0.6): cadastral sections colored by median price, clickable for details
- Price stats: average price/m², median price/m², number of sales
- Apartments / Houses toggle
- Comparison mode: compare two arrondissements side-by-side with a bar chart
- Address search (v0.5): search by address and get stats for the cadastral section
- UI-driven search (v0.8): search addresses and browse arrondissements directly from the widget
Data source: DVF (Demandes de Valeurs Foncieres) from data.gouv.fr.
Architecture
Claude MCP Server UI (iframe)
│ │ │
│── "prix Paris 11" ──────>│ │
│ │── get-dvf-stats ───────>│
│ │ structuredContent │── Map + Stats
│ │ │
│── "compare 6e vs 11e" ──>│ │
│ │── compare-dvf-stats ───>│
│ │ mode: "compare" │── Map + Bar chart
│ │ │
│── "prix rue Roquette" ──>│ │
│ │── search-dvf-address ──>│
│ │ mode: "address" │── Map + Marker + Compare
- Claude calls
get-dvf-stats,compare-dvf-statsorsearch-dvf-address - The tool returns data + a reference to
ui://dvf/mcp-app.html - Claude fetches the resource and displays it in a sandboxed iframe
- The UI receives data via
app.ontoolresult
MCP Tools
| Tool | Input | Output |
|---|---|---|
get-dvf-stats |
arrondissement (1-20) |
Price stats for one arrondissement |
compare-dvf-stats |
arrondissement_1, arrondissement_2 (1-20) |
Side-by-side comparison with bar chart |
search-dvf-address |
adresse (string) |
Stats for cadastral section + comparison with arrondissement |
Stack
| Component | Technology |
|---|---|
| MCP Apps SDK | @modelcontextprotocol/ext-apps |
| MCP Server | @modelcontextprotocol/sdk |
| Build | Vite + TypeScript |
| UI | Vanilla JS |
| Map | Leaflet + OpenStreetMap (no API key needed) |
| Charts | Pure SVG (no dependencies) |
| Transport | stdio (Claude Desktop) or Streamable HTTP |
External APIs
| API | Usage |
|---|---|
| MCP data.gouv | Real-time DVF stats |
| API Géoplateforme | Address geocoding |
| cadastre.data.gouv.fr | Cadastral section geometries (GeoJSON) |
Setup
npm install
npm run build
Claude Desktop configuration
Add to ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"dvf-paris": {
"command": "bash",
"args": ["-c", "cd /path/to/dvf-mcp-app && npx tsx main.ts --stdio"]
}
}
}
Then restart Claude Desktop.
Development
# Watch mode (UI hot reload + server)
npm run dev
# Build only
npm run build
# Run server (Streamable HTTP on port 3001)
npm run serve
Project structure
dvf-mcp-app/
├── server.ts # MCP Server + Tools + Resource
├── main.ts # Entry point (stdio + HTTP)
├── mcp-app.html # UI shell (HTML)
├── src/
│ ├── mcp-app.ts # UI logic (map, chart, host communication)
│ ├── mcp-app.css # Widget styles
│ ├── api/ # External API clients
│ │ ├── cadastre.ts # Cadastre GeoJSON client
│ │ ├── data-gouv.ts # MCP data.gouv client
│ │ └── geoplateforme.ts # Geocoding client
│ └── data/
│ ├── dvf-paris.json # Pre-computed stats (fallback)
│ └── arrondissements.geojson.json # GeoJSON boundaries
├── package.json
├── tsconfig.json
└── vite.config.ts
Roadmap
- [x] v0.1 — Basic stats widget
- [x] v0.2 — Interactive map with Leaflet
- [x] v0.3 — Comparison mode (2 arrondissements)
- [x] v0.4 — Real-time data via data.gouv.fr API (with JSON fallback)
- [x] v0.5 — Address search with cadastral section stats
- [x] v0.6 — Choropleth cadastral sections (clickable, color-coded by price)
- [ ] v0.7 — Link to recent transactions (optional)
- [x] v0.8 — Fullscreen mode with interactive search (split layout, callServerTool)
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.