What we publish, and what we will not

Our own aggregates. Medians, percentiles, distributions, counts, shares, rankings, histograms, pipelines and comparisons that we computed. These did not exist before we computed them. They are our work, and they are what the free pages and the paid products consist of.

Developer-published project facts. The name of a project, who is developing it, where it is, what property types and bedroom counts it contains, its construction and sale status, its scheduled handover and the structure of its published payment plan. These are facts that a developer states publicly about its own product in order to sell it. Facts are not owned by whoever repeats them, and we credit the developer as their source.

In this build that comes to 2,924 off-plan projects and 52,882 priced unit records reduced into aggregates across 230 communities and 676 developers. How that reduction works is set out in full on the methodology page.

• Portal listing records, verbatim or near-verbatim. No individual listing as it appears on PropertyFinder, Bayut, Reelly or anywhere else. Not a reformatted copy of one, and not a bulk export of them. What we publish is the statistic computed across many of them.
• Portal images. No renders, floor plans, photographs or brochure pages taken from a portal are hosted, hotlinked or embedded on this site or in any paid product. Every image we do use is either generated by us or licensed by us.
• Portal descriptions. No marketing copy, project write-up, amenity blurb or agent description is reproduced. Where a page needs prose, we write it.
• Portal URLs. We do not publish deep links into listing pages, and we do not pass listing identifiers through into our exports. The pipeline strips source URLs, brochure URLs and image references before anything reaches the site bundle, so they are not merely hidden, they are absent.
• Anything obtained from behind a login. Nothing in this project authenticates against a portal, and nothing here was taken from a subscriber-only area.

The portals we read are not confined to the UAE, and not every row they return is a project. Both are filtered out in the pipeline, before any median, count or benchmark on this site is computed, so a page can never quote a figure that a discarded record helped produce.

• Anything outside the UAE. Reelly lists Bali, Oman and Turkey next to Dubai. On the corpus as it stood when this rule was written, 79 such projects, holding 1,369 priced units, were being counted in numbers this site labelled UAE. They pulled the UAE median price per sqft down to AED 1,950 from the AED 1,974 the rest of the book was asking at the time, and the median ticket to AED 1.86m from AED 1.90m. A project is kept only if it sits within 25 km of a project a portal has itself labelled with an emirate; the nearest record the rule rejects is a Muscat tower 270 km out.
• Rows carrying no data. 148 records arrived with a title and nothing else: no developer, no coordinates, no units, no price, no handover date. 80 duplicated a project already held under a proper address. They moved no median, having no number to move one with, but they were 3.4 percent of the project count printed on this site.

That is 227 of 4,315 records discarded at the time, leaving 4,088. A later pass merged 96 duplicate groups, ingested what the watcher has found since, and split the book into 2,924 off-plan projects and 1,087 already-delivered comparables, 4,011 records in all. The same pass reads the emirate for a project the portal left unlabelled from the projects it did label, which recovered 35 real Abu Dhabi developments, 183 priced units among them, that a previous version of the rule had placed nowhere at all.

We credit the developer as the source of the project facts, because that is where those facts originate. We name PropertyFinder, Bayut and Reelly as the public portals where the listings were observed, which is a statement about where we looked, not a claim of partnership, endorsement or licence. None of them sponsors, reviews or has any involvement in this site. All trade marks belong to their owners.

Any figure you see is attributable to OffplanIndex. If you quote one, quote it as ours and say which build it came from, because the numbers move between builds. Sources were read to 2026-07-30, the corpus was merged on 2026-08-18, and these tables were computed on 2026-09-02.

Here is our reasoning, in the same words we would use if you asked us over coffee.

A database can be protected. The individual facts inside it usually cannot be, because a fact about the world belongs to nobody. The price a developer is asking for a two bedroom apartment is a fact about that developer’s product. Observing many such facts and computing what the middle of the market looks like creates something new, and that new thing is ours, in the same way a stock index is the index provider’s even though every price in it came from an exchange.

What that reasoning does not cover is taking the shape of somebody else’s collection: their selection, their arrangement, their images, their prose, their identifiers, their record boundaries. So we do not take any of it. We keep no copy of a portal database, we publish nothing from which one could be reconstructed, and our exports are deliberately built so that a reader gets the statistic and never the underlying rows.

If you believe we have got that balance wrong on a specific page or a specific figure, we would rather hear it than argue about it. We will look at the actual page, and if you are right we will change or remove it.

Write to hello@offplanindex.com. Tell us the exact URL, what specifically is on it that you object to or believe is wrong, and who you are. There is no form and no ticket system on purpose.

We aim to acknowledge within two working days. Where the complaint concerns republished content of yours, which should not exist given the rules above, we will remove it first and discuss afterwards. Where it concerns a figure you believe is inaccurate, we will re-run the query and either correct it in the next build or show you the records behind it.

You may read these pages, quote a figure in an article, a valuation, a client presentation or a social post, and link to any page here. When you quote a figure, attribute it to OffplanIndex with a link to the page it came from, and say which build it is from. A journalist, an agent or an analyst using our numbers to make a point is exactly what these pages are for.

You may not scrape these pages in bulk, systematically extract the tables, mirror the site, or build a competing dataset or product out of the aggregates. Automated bulk retrieval is not permitted regardless of the rate at which it is done. If you want the tables as data rather than as pages, that is a product, and it is cheaper than building a scraper: see the report or write to us about a data licence.

The report, including its CSV. Sold for use by the single buyer, meaning one person or one company that bought it. Use it internally as much as you like: analyse it, chart from it, quote it in your own client work with attribution. Do not redistribute the file, resell it, post it publicly, or pass it to another firm. A team that all needs it should buy the seats or ask us about a company licence.

The agent seat. One seat is one person. Credentials are not shared between colleagues, and an office does not run on one login. Client-ready one-pagers generated from a seat are yours to send to your clients, brand as your own and use commercially, which is what the seat is for. The underlying data behind them stays with the seat holder, and bulk export of inventory out of the app is not part of the seat.

Both products are covered by the terms. Nothing sold here is investment advice.