Darak MCP Server
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Darak MCP Servershow me apartments for rent in Riyadh under 5000 SAR"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Darak MCP Server
Remote MCP server for Saudi real estate data. Gives AI assistants access to 65,000+ rental and sale property listings across 5 Saudi cities, with market analytics, neighborhood comparisons, and price trends.
Server URL: https://platform.darak.app/mcp
Tools (23, all read-only)
Search & Listings
Tool | Description |
| Search rental and sale listings with filters |
| Fetch several listings by ID |
| Count listings matching search filters |
| Full details for one listing |
| Similar nearby listings for price comparison |
| Price changes over time for a listing |
| Listings priced below their neighborhood median |
Market Analytics
Tool | Description |
| Price histogram and percentiles |
| Area (sqm) distribution |
| Listing price and area context |
| Side-by-side neighborhood comparison |
| Gross yield from rental and sale medians |
| Listing supply over time |
| Stale listing counts |
| Neighborhood rent values for map visualizations |
| City-level market overview |
| Monthly neighborhood price trends |
Geography
Tool | Description |
| Neighborhood IDs and names by city |
| City districts and their neighborhoods |
Off-plan Projects
Tool | Description |
| Find developments by location and filters |
| Project details |
| Developers with off-plan projects |
| Find available units in projects |
Related MCP server: nj-realtor-mcp
Cities
Riyadh, Jeddah, Eastern Province, Makkah, Madinah.
Connect
Claude Desktop / Claude Code
{
"mcpServers": {
"darak": {
"command": "npx",
"args": ["mcp-remote", "https://platform.darak.app/mcp"]
}
}
}Claude Code (CLI)
claude mcp add --transport http darak https://platform.darak.app/mcpAny MCP client (Streamable HTTP)
Connect directly to https://platform.darak.app/mcp using the Streamable HTTP transport.
Connected accounts and OAuth
Anonymous access uses a shared service key and a per-caller budget. When that
budget is exhausted, the server returns a Bearer challenge pointing to
https://platform.darak.app/.well-known/oauth-protected-resource. Its metadata
names https://platform.darak.app/mcp as the resource, darak.read as the scope, and
https://darak.app/api/auth as the authorization-server issuer. Production
BETTER_AUTH_URL uses darak.app; even metadata fetched from
platform.darak.app advertises the darak.app/api/auth issuer. Clients discover
the authorize/token/registration endpoints from that issuer;
the Worker is a resource server and does not mint credentials.
A connected caller's Bearer token is forwarded unchanged to the versioned Darak API, which validates the issuer, audience, expiry, and scope and meters the caller's organization. DPoP and malformed authorization headers are rejected, not silently treated as anonymous; DPoP-bound tokens sent as Bearer are rejected by the API. The Worker cannot support DPoP-bound access tokens without verifying proofs itself or coordinating a proof-preserving API call.
Hard cutover: The old https://darak.app/mcp route will be retired, not
redirected. Update every MCP client configuration to https://platform.darak.app/mcp
and reconnect: old-resource tokens and consent do not transfer. Keep the OAuth
issuer at https://darak.app/api/auth. Deploy the app's guarded resource migration
and new audience verification before switching this Cloudflare Worker route.
After deployment, test discovery, S256 PKCE, consent, token refresh, and a
connected MCP call. Clients may cache old metadata for up to an hour.
Development
npm install
npm run dev # Local dev server at http://localhost:8787
npm run type-check
npm test # Node 24
npm run deploy # Deploy only after the coordinated rolloutArchitecture
Runs on Cloudflare Workers with Durable Objects
Calls the public, versioned, metered Darak API at
https://api.darak.app/v1All tools are read-only (annotated with
readOnlyHint: true)Anonymous use is available; connected accounts can use OAuth Bearer tokens
Privacy Policy
See https://darak.app/privacy for the full privacy policy.
Data handling summary:
Authentication is optional. A connected caller's access token is forwarded to the Darak API for authorization and metering; this Worker code does not mint or persist it. Anonymous usage budgets are keyed by a salted hash of the caller IP.
No conversation data stored. Queries are proxied to the Darak API and responses are returned directly. The server does not log, store, or inspect query contents.
Anonymous usage analytics. Tool call events (tool name, city, listing type, success/failure) are sent to PostHog for aggregate usage monitoring. No personally identifiable information is included.
No third-party data sharing. Data is not sold, shared, or transferred to third parties beyond the PostHog analytics described above.
Data source. All property data is aggregated from publicly available Saudi real estate platforms.
Support
This server cannot be deployed
Maintenance
Related MCP Connectors
MCP server giving Claude AI access to 22+ NYC public-record databases for real estate due diligence
- UnifAPIOAuthcom.unifapi
Hosted MCP server for live public-data APIs and Skills for AI agents.
Hosted MCP server exposing US hospital procedure cost data to AI assistants
Remote MCP server exposing 330 production AI-agent services for web/data processing, validation, AI utilities, blockchain/crypto utilities, and x402 pay-per-use access.
Related MCP Servers
- AlicenseAqualityCmaintenanceEnables AI-powered real estate analysis with built-in EU AI Act compliance, providing a production-ready MCP server for property insights and governance.1MIT
- FlicenseNot gradedqualityDmaintenanceA production-grade MCP server enabling Claude to perform comprehensive NJ real estate workflows including property search, valuation, neighborhood intelligence, investment analysis, and agent tools via 20 tools and 15+ data sources.-
- AlicenseNot gradedqualityDmaintenanceProvides comprehensive real estate market intelligence, property valuation, and investment analysis for AI agents. Unifies data from multiple sources like Zillow, Redfin, and public records into a single MCP interface.MIT
- AlicenseNot gradedqualityDmaintenanceAn MCP server that gives AI assistants access to a multi-agent business operations platform with specialized AI services covering real estate, legal, translation, research, content creation, training, and more.36 npmMIT