DEV Community

Daemon Data
Daemon Data

Posted on

Is there a Dubai Land Department (DLD) API? Registered sales, Ejari rents and projects as JSON in Python

Short answer: not a developer API with keys and documentation, but the data is open and one POST away. The Dubai Land Department (DLD) publishes its registers as open data on its Real Estate Data page: every sale, mortgage and gift it registers, every Ejari rent contract, and the register of projects and developers. The page offers a search form and a CSV download. Behind the form is a JSON gateway, and it answers the same request from a script. On 17 November 2026 one request told me that DLD registered 17,648 transactions in October 2026, and a short loop returned all 749 transactions of 12 November.

This post shows the calls for sales and rents with Python that ran on 17 November 2026, how to page without losing rows, and the limits you should know before you build on it, the largest being that the gateway serves the current calendar year only.

What DLD offers officially

  • The open-data pages on dubailand.gov.ae: search by date and filters, CSV export of what the search returns. This is the source used below.
  • Data.Dubai, the emirate's data portal, lists a "Real Estate Transactions" dataset with the earlier years. When I tried it from outside the UAE on 10 November 2026 I could not download the files.
  • Paid products: property portals and data vendors sell cleaned DLD data with their own analysis, by subscription.

There is no API key to request for the open data, and no published rate limit. Treat it as a public service: one request at a time, with a pause.

Step 1: one request, the register for a date range

Dates are MM/DD/YYYY. Every filter must be present in the body; an empty string means "all".

import requests

GATEWAY = "https://gateway.dubailand.gov.ae/open-data/"
HEADERS = {
    "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 "
                  "(KHTML, like Gecko) Chrome/141.0.0.0 Safari/537.36",
    "Accept": "application/json, text/plain, */*",
    "Origin": "https://dubailand.gov.ae",
    "Referer": "https://dubailand.gov.ae/",
}


def transactions(date_from, date_to, take=100, skip=0):
    """One page of DLD transactions registered between two dates (MM/DD/YYYY, both included)."""
    body = {
        "P_FROM_DATE": date_from, "P_TO_DATE": date_to,
        "P_GROUP_ID": "",        # "" all, 1 sales, 2 mortgages, 3 gifts
        "P_IS_OFFPLAN": "", "P_IS_FREE_HOLD": "", "P_AREA_ID": "",
        "P_USAGE_ID": "",        # 1 residential, 2 commercial
        "P_PROP_TYPE_ID": "",    # 1 land, 2 building, 3 unit
        "P_SORT": "TRANSACTION_NUMBER_ASC",
        "P_TAKE": str(take), "P_SKIP": str(skip),
    }
    r = requests.post(GATEWAY + "transactions", json=body, headers=HEADERS, timeout=120)
    r.raise_for_status()
    return r.json()["response"]["result"]


rows = transactions("10/01/2026", "10/31/2026", take=3)
print(rows[0]["TOTAL"], "transactions registered in October 2026")
for t in rows:
    print(t["INSTANCE_DATE"], "|", t["GROUP_EN"], "|", t["PROCEDURE_EN"], "|", t["AREA_EN"], "|",
          t["PROP_TYPE_EN"], t["PROP_SB_TYPE_EN"], "|", t["ACTUAL_AREA"], "sqm | AED", t["TRANS_VALUE"])
Enter fullscreen mode Exit fullscreen mode
17648 transactions registered in October 2026
2026-10-23T08:43:58 | Mortgage | Portfolio Mortgage Registration Pre-Registration | HORIZON | Unit Office | 459.56 sqm | AED 15300000
2026-10-23T08:43:58 | Mortgage | Portfolio Mortgage Registration Pre-Registration | HORIZON | Unit Office | 320.79 sqm | AED 15300000
2026-10-02T10:59:26 | Sales | Sell - Pre registration | DUBAI PRODUCTION CITY | Unit Flat | 77.33 sqm | AED 924137.42
Enter fullscreen mode Exit fullscreen mode

Every row carries TOTAL, the number of rows the whole query has, so the first request tells you the size of the job. The fields you will use most:

Field Meaning
TRANSACTION_NUMBER DLD's number of the transaction; several rows can share it
INSTANCE_DATE when it was registered, Dubai time
GROUP_EN, PROCEDURE_EN Sales, Mortgage or Gifts, and the procedure (for example Sell - Pre registration for an off-plan sale)
IS_OFFPLAN_EN, IS_FREE_HOLD_EN off-plan or ready, freehold or not
AREA_EN, PROJECT_EN, MASTER_PROJECT_EN where
PROP_TYPE_EN, PROP_SB_TYPE_EN, ROOMS_EN land, building or unit; flat, villa, office; rooms
TRANS_VALUE value in AED
ACTUAL_AREA, PROCEDURE_AREA square metres
NEAREST_METRO_EN, NEAREST_MALL_EN, NEAREST_LANDMARK_EN DLD's own location hints

Arabic versions of the text fields end in _AR.

Step 2: every row, without repeating or losing any

P_TAKE and P_SKIP page the result. The trap: DLD sorts by transaction number only, and a portfolio mortgage puts dozens of rows under one number. The gateway orders the rows inside such a group differently on every request, so when a page boundary cuts a group, the next page repeats some of its rows and drops others. In October 2026 a plain 500-row loop repeated 13 rows and lost 7.

The fix is to never split a group: drop the trailing group from each page and start the next page at its first row.

import time


def all_transactions(date_from, date_to, take=2000):
    """Every transaction of the range. A group of rows with one transaction number is never split across pages."""
    out, skip = [], 0
    while True:
        page = transactions(date_from, date_to, take, skip)
        if not page:
            return out
        last = skip + len(page) >= page[0]["TOTAL"]
        keep = len(page)
        if not last:
            tie = page[-1]["TRANSACTION_NUMBER"]
            while keep > 0 and page[keep - 1]["TRANSACTION_NUMBER"] == tie:
                keep -= 1
            if keep == 0:           # one transaction fills the page: ask for a bigger one
                take *= 2
                continue
        out += page[:keep]
        skip += keep
        if last:
            return out
        time.sleep(1.5)


day = all_transactions("11/12/2026", "11/12/2026")
sales = [t for t in day if t["GROUP_EN"] == "Sales" and t["PROP_TYPE_EN"] == "Unit" and t["ACTUAL_AREA"]]
per_sqm = sorted(t["TRANS_VALUE"] / t["ACTUAL_AREA"] for t in sales)
print(len(day), "transactions on 12 November 2026;", len(sales), "unit sales; median AED",
      round(per_sqm[len(per_sqm) // 2]), "per sqm")
Enter fullscreen mode Exit fullscreen mode
749 transactions on 12 November 2026; 382 unit sales; median AED 18248 per sqm
Enter fullscreen mode Exit fullscreen mode

With this loop I read the whole of 2026 to 16 November month by month: 190,347 rows, and every month's count equalled the gateway's own total.

Step 3: Ejari rent contracts

Rent contracts registered in Ejari are a second endpoint with the same shape and two differences: there is a date-type parameter, and the To date is exclusive.

def rents(date_from, date_to_exclusive, take=5000, skip=0):
    """One page of Ejari contracts registered from date_from up to, not including, date_to_exclusive."""
    body = {
        "P_DATE_TYPE": "0",      # 0 registration date
        "P_FROM_DATE": date_from, "P_TO_DATE": date_to_exclusive,
        "P_VERSION": "",         # "" all; new and renewed contracts can be asked separately
        "P_IS_FREE_HOLD": "", "P_USAGE_ID": "", "P_AREA_ID": "", "P_PROP_TYPE_ID": "",
        "P_SORT": "REGISTRATION_DATE_ASC",
        "P_TAKE": str(take), "P_SKIP": str(skip),
    }
    r = requests.post(GATEWAY + "rents", json=body, headers=HEADERS, timeout=240)
    r.raise_for_status()
    return r.json()["response"]["result"]


contracts = rents("11/12/2026", "11/13/2026", take=3)
print(contracts[0]["TOTAL"], "rent contracts registered on 12 November 2026")
for c in contracts:
    print(c["REGISTRATION_DATE"], "|", c["VERSION_EN"], "|", c["AREA_EN"], "|", c["PROP_TYPE_EN"], c["PROP_SUB_TYPE_EN"],
          "|", c["ACTUAL_AREA"], "sqm | AED", c["ANNUAL_AMOUNT"], "a year | properties in contract:", c["TOTAL_PROPERTIES"])
Enter fullscreen mode Exit fullscreen mode
4851 rent contracts registered on 12 November 2026
2026-11-12T00:18:14 | Renewed | Business Bay | Unit Flat | 146 sqm | AED 200000 a year | properties in contract: 1
2026-11-12T00:19:03 | Renewed | Marsa Dubai | Unit Flat | 81 sqm | AED 77000 a year | properties in contract: 1
2026-11-12T00:43:03 | Renewed | Hadaeq Sheikh Mohammed Bin Rashid | Unit Flat | 145 sqm | AED 210000 a year | properties in contract: 1
Enter fullscreen mode Exit fullscreen mode

Rents are the bigger register: about 3,000 to 5,000 contracts on a working day, 95,572 in October 2026. Page them the same way, keeping rows with the same REGISTRATION_DATE second together.

The gotchas

  1. Current year only. The same call for October 2025 returns nothing:
   print(len(transactions("10/01/2025", "10/31/2025", take=3)), "rows for October 2025")
Enter fullscreen mode Exit fullscreen mode
   0 rows for October 2025
Enter fullscreen mode Exit fullscreen mode

If the gateway treats 2026 the way it treats 2025 today, the 2026 rows stop being served on 1 January 2027. If you will need 2026, read it before the year ends.

  1. A portfolio mortgage repeats its value. The first two rows above are two offices under one mortgage; each carries the whole AED 15.3 million. Sum TRANS_VALUE over mortgage rows and you count it once per unit. Sales rows are one property each.
  2. Dates are MM/DD/YYYY in the request and ISO timestamps in the answer.
  3. Paging needs the group rule from step 2, or totals drift by a few rows a month.
  4. The register changes behind you. Rows are added to recent days for a while; October's total was 17,646 when I first read it in early November and 17,648 on 17 November. Re-read the last week or two when you update.
  5. USAGE_EN is empty. Since 9 November 2026 the gateway returns no usage name on sales rows; USAGE_ID still says 1 for residential and 2 for commercial.
  6. Rents: the To date is exclusive, and a contract that covers several properties appears once per property with the whole contract amount on each row (TOTAL_PROPERTIES tells you). The property-type filter on rents returned no rows for any value when I tested it on 7 November; filter on PROP_TYPE_EN yourself.
  7. Areas are square metres, values are dirhams, and the gateway has no price per square metre: divide.

DLD also serves the register of projects and developers the same way (915 projects when I read it on 9 November); that is a post of its own.

Abu Dhabi is a different source

Abu Dhabi's sales are published by the Abu Dhabi Real Estate Centre (ADREC) on its market-data page, and unlike Dubai's they go back to January 2019: 127,379 sales when I counted on 12 November 2026. The numbers by district are in What do homes actually sell for in Abu Dhabi?

If you would rather not maintain it

I maintain the calls above as hosted Actors on Apify: Dubai DLD Property Transactions (sales, mortgages and gifts with a computed price per square metre), Dubai Rent Contracts (Ejari) and Abu Dhabi Property Transactions. Each can back-fill a date range or run on a schedule and deliver each row once.

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("data_daemon/dubai-property-transactions").call(run_input={
    "daysBack": 7,
    "maxItems": 200,
})
for row in client.dataset(run["defaultDatasetId"]).iterate_items():
    if row["group"] == "Sales":
        print(row["transactionDate"], row["area"], row["propertySubType"], row["rooms"], row["actualAreaSqm"],
              row["transactionValueAed"], row["pricePerSqmAed"], "off-plan" if row["offPlan"] else "ready")
Enter fullscreen mode Exit fullscreen mode
2026-11-12 THE WORLD Villa 6 B/R 2274.76 60900000 26772.06 off-plan
2026-11-11 Jabal Ali First Flat 2 B/R 107.15 1437000 13411.11 off-plan
2026-11-16 Dubai Investment Park Second Villa 4 B/R 265.3 3300000 12438.75 off-plan
Enter fullscreen mode Exit fullscreen mode

Those are the first three of 162 sales among the 200 rows of that run. Dubai and Abu Dhabi sales cost $1 per 1,000 rows until 28 November 2026 and $2 from 29 November, so the whole of 2026 to date (190,347 Dubai transactions) is about $190 now and about $380 after; one month is about $18, then $36. Ejari rents are $1 per 1,000 and do not change.

Data is the Dubai Land Department's public register, shown for information; it is not valuation or investment advice.

Top comments (0)