Build with Darak and AI
Choose the integration that matches who is making the request:
| Need | Use | Credentials |
|---|---|---|
| A backend, scheduled job, or product that syncs or analyzes Darak data | REST API or JavaScript/Python SDK | A server-side data key; use an admin key only for organization management |
| A person asking an AI assistant about listings or markets | Darak MCP | The person signs in with their Darak account; calls use their organization's plan |
| Guidance for an assistant to combine several Darak tools | Optional Darak skill | The skill provides instructions, not a credential or an MCP connection |
An MCP client can answer "Find 2-bedroom rentals in Riyadh's Al Olaya and tell me whether each is expensive" by calling list_neighborhoods for the neighborhood ID, search_listings for candidates, and get_listing_market_stats for pricing context. For investment questions, it can compare get_rental_yield and compare_neighborhoods. These are read-only tools, not a way to create a listing or administer an organization.
All listing prices in Darak are normalized to SAR per year; an advertiser's monthly or daily amount is also preserved separately. A neighborhood name is not an ID: look it up before filtering. Comparable prices and yields are estimates from the available inventory, not appraisals or guaranteed returns. Keep provenance and listing freshness visible when you present AI-generated answers.
For deterministic product workflows, call the versioned API from your backend instead of scraping or asking an assistant to paginate indefinitely. The API quickstart shows curl and both SDKs; coverage and freshness explains the data's limits. If you install the optional skill, point it at the same canonical MCP endpoint and review its advice against the live tool list.