The six checks, and what normal looks like
| Check | Normal across 966 registrations |
|---|---|
| 1. Find the registration | The number on the filing equals the number on your paperwork |
| 2. Read the promoter of record | A named company, and 861 of 966 filings are companies |
| 3. Count the documents | Median 21, quartiles 16 and 29 |
| 4. Compare the completion period proposed with the date now showing | Median period 60 months; 519 of 957 have moved the date |
| 5. Read the extensions as a chain | Median 6 months; 526 of 966 carry one |
| 6. Read the complaints against the filing's own cohort | 87 percent of 2017 to 2019 filings carry one, 14 percent of 2025 to 2026 |
Why a filing is unreadable without denominators
Opening a UP-RERA registration is easy and free. Reading it is not, because every field on it is a number without a scale. Twenty-one documents: is that thorough or thin? A sixty-month completion period: generous or tight? Two extensions: routine or alarming? No complaints: clean, or just new?
Nothing published anywhere answered those questions, so we counted all 966 Gautam Buddha Nagar registrations on our mirror of the register, read 29 July 2026, and published each answer. This page is the procedure those numbers were for. It is the sequence our desk runs before a client signs, and there is nothing in it you need us for.
1. Find the registration, and check the number is the one on your paperwork
Search the project name on the authority's site, not the promoter's. Then read the registration number on the filing against the number printed on your own allotment letter or agreement, character for character. Brand names are shared here: a single brand word can cover several different registered companies, which is why the number rather than the name is the thing that identifies what you are buying.
The register is at up-rera.in and the same records are browsable at upreratracker.com. Full read: what a normal filing looks like.
2. Read the promoter of record, and its type
The promoter named on the filing is the company that owes you the home, and it is often not the company on the hoarding. Check the name word for word against your agreement. The applicant type tells you what kind of entity it is: of 966 registrations, 861 are filed by a Company, 46 by a Partnership Firm, 18 by an LLP and 27 by a development authority itself.
479 distinct promoters hold the 966 registrations. Full read: who registers projects here.
3. Count the documents
The filing lists what the promoter uploaded. The district median is 21, with quartiles 16 and 29 and a maximum of 108. A count in the teens is ordinary and a count below ten is worth a question, though a high count is completeness rather than quality. Open the sanctioned building plan and the encumbrance details if they are there.
22986 documents across the register, in 25 types. Full read: the document counts in full.
4. Compare the completion period proposed with the date now showing
Two dates matter: the completion the promoter declared at registration, and the date the filing currently carries. The median proposed period here is 60 months. Of 957 filings whose dates both parse, 519 now show a later date than they declared and 438 still show the original. The median slip is 6 months.
Of the 704 filings from 2017 to 2019, 503 moved the date. Full read: completion dates across 957 filings.
5. Read the extensions as a chain, not as events
An extension is a revised date entered on the register, and 526 of the 966 registrations carry at least one. The median length is 6 months. Where a filing carries several, read the date at the end of the run rather than the length of any one grant: on 198 of the 206 registrations with two or more, each begins the day the last one ended.
455 records state a reason and every one is the covid force majeure allowance. Full read: all 852 extension records.
6. Read the complaints against the filing's own cohort
This is the field people misread hardest, in both directions. 617 of 711 registrations from 2017 to 2019 carry at least one complaint, against 14 of 97 from 2025 to 2026. So a clean column on a 2026 filing is the ordinary case and tells you little, while a clean column on a 2017 filing is genuinely unusual. A complaint is an allegation filed, not a finding against anyone; if the count is high for its cohort, read what the orders say it was about.
Across those that carry any, the median is 15. Full read: the complaints read in full.
What a completely clean filing looks like, and how rare it is
Put the three moving columns together, and clean means no extension, no complaint, and a completion date that was never revised. Counted 23 September 2026: 120 of the 148 registrations filed in 2024 or later are clean on all three, about 81 percent. Of the 818 filed before 2024, 87 are, about 11 percent.
That single comparison is the most useful thing on this page. On a recent registration, clean is close to normal and proves very little. On an older one, clean is genuinely rare and means something. Any read of these fields that ignores the registration date is reading the age of the filing and calling it the quality of the promoter.
What none of this establishes
Not build quality, not financial standing, not whether the promoter will still exist in five years. The register records filings, dates, documents, and allegations made; it does not inspect anything and it does not decide anything. These six checks tell you whether what you have been told matches what has been filed, and they tell you which questions are worth asking next. That is a great deal more than most buyers have when they pay a booking amount, and it is still not a verdict.
Everything above is reproducible from the public record. The authority is up-rera.in and the same registrations are browsable at upreratracker.com. Where a brand name appears under several registered companies, which is common here, our read of all 966 filings covers how to tell them apart.
