Tamleek Plus — how to use it
Tamleek Plus is a real-estate intelligence platform for the Qatar market. It combines Ministry of Justice (MoJ) registered sale data, Aqarat QREP rental/mortgage data, and active listing data (currently The Pearl, pilot scope) into indicative pricing, trend, and valuation signals.
There are three ways to use it, depending on who you are and what you're building:
| Interface | Who it's for | What you get |
|---|---|---|
| MCP | AI agents / IDE tools (Claude, Cursor, etc.) | 10 deterministic tools your own agent can call directly |
| REST API | Your own backend / a script | The same data over plain HTTP, JSON in/out |
| Platform UI | A person, in a browser | A chat analyst + a map, no code required |
One data model, three doors
All three interfaces sit on top of the same underlying data and the same rules about what counts as a fact. That means:
- MoJ registered deed values are the source of truth for what something actually sold for. Attribute these as "MoJ registered deed values" — never as "the price" without qualification.
- Aqarat rental/mortgage data covers rent and financing activity — never presented as sale prices.
- Active listing asking prices (currently The Pearl only) are a supply/demand signal, not a sold price. They are always labelled as asking prices.
- The Pearl synthetic demo (a small set of demo towers) is entirely simulated data used for AVM methodology testing with UDC — never a real transaction, real asking price, or a valuation. Every response using it says so.
Every response, on every interface, carries a disclaimer: what the number is (registered / asking / synthetic), how many comparable records back it, and how fresh the data is. None of the three interfaces produce an investor-grade valuation — everything is indicative.
Compliance boundary that applies everywhere
Qatar's Personal Data Privacy Protection Law (PDPPL, Law 13/2016) means no interface exposes unit-level, individually-identifiable transaction data (buyer/seller names, specific-unit deed records for a named private owner). Every surface answers at the district/precinct/tower level, or from an explicitly-synthetic demo dataset labelled as such. If you need unit-level detail for a specific transaction, that is out of scope for all three interfaces.
Where to go next
- Building an AI agent or wiring this into Claude/Cursor? Start with MCP.
- Building your own app or script against the data? Start with the API.
- Just want to ask questions and see a map? Use the platform UI.