Name Match API: is the patta holder the same person as the seller?
Yes. name_match compares two names as they are written on two records, such as the patta holder and the seller in a sale deed, or an EC party and an uploaded document, and says whether they are the same person: match, partial_match (a relative or one of several parties), cannot_tell, or no_match, with a confidence from 0 to 100 and the reason. It reconciles initials and name order (S. Raman and Raman S), spelling variants, and Tamil and English forms of one name. It is an advisory reading of two strings, not proof of identity. It costs 1 credit per comparison.
https://verified.realestate/api/v1/tools/name_match- Price
- 1 credit per answer
- Source
- A language model reading two names
- Coverage
- Tamil and English names
- Answers in
- Usually 15 to 25 seconds
What it returns
One pair, one verdict.
| Field | What it holds |
|---|---|
verdict, verdict_label | match, partial_match, cannot_tell or no_match, and the same in words. |
match, partial_match | The verdict as two flags. |
confidence | 0 to 100. Under 70, treat the answer as unverified. |
reason | Why, in a sentence or two. |
relationship | How the two relate: same_person_unconfirmed, family, group_party (one of several parties), entity_variant (a company written two ways) or unrelated. |
what_would_settle_it | On cannot_tell: the document or fact that would decide it. |
name_a, name_b, compared | What was compared, and the label you sent in context. |
advisory, identity_proof_required | Always true: the answer does not prove identity. |
Inputs
The JSON body of the POST is the arguments. Both names are required. Send them as written, not cleaned up.
| Argument | Type | What to send |
|---|---|---|
name_arequired |
string | The first name exactly as recorded, Tamil script included, e.g. the patta holder. |
name_brequired |
string | The second name exactly as recorded, e.g. the seller in the sale deed. |
context |
string | Optional. What the two names are, e.g. "patta vs sale deed". The reason then names your two documents. |
Code
Set LANDLENS_KEY to your key. In an AI app, connect https://verified.realestate/mcp once and ask in plain words instead. In Postman, download this API's collection, import it (File, Import) and set apiKey.
curl -X POST https://verified.realestate/api/v1/tools/name_match \
-H "Authorization: Bearer $LANDLENS_KEY" \
-H "Content-Type: application/json" \
-d '{}'
import os, requests
r = requests.post(
"https://verified.realestate/api/v1/tools/name_match",
headers={"Authorization": f"Bearer {os.environ['LANDLENS_KEY']}"},
json={},
timeout=60,
)
body = r.json()
print(body["status"], body["credits_charged"], body["data"])
const res = await fetch("https://verified.realestate/api/v1/tools/name_match", {
method: "POST",
headers: {
Authorization: `Bearer ${process.env.LANDLENS_KEY}`,
"Content-Type": "application/json",
},
body: JSON.stringify({}),
});
const body = await res.json();
console.log(body.status, body.credits_charged, body.data);
The patta says the holder is R. Kumaresan and the sale deed seller is Kumaresan s/o Ramasamy. Is that the same person?
Your AI says the price, compares the two names, and tells you what document would settle any doubt.
Responses
Every answer has the same envelope. What status can say for this call:
| status | What it means |
|---|---|
ok | The comparison ran and a verdict is given. Charged 1 credit. |
not_found | The comparison could not run. The names are not reported as different. Nothing is charged; try again. |
invalid_args | name_a or name_b is missing. |
Where the data comes from
No register. The two names you send are compared by a language model instructed in Tamil Nadu naming conventions: initials, father's and husband's names, caste suffixes, and Tamil and English spellings.
Any two names. Built for the names on Tamil Nadu land records, in Tamil or English script.
Not applicable. Each call compares the two names you send.
The model returns whether the names match, partially match, or do not, with a confidence and a reason. A partial match with low confidence, or one the model marks as possibly the same person, is reported as cannot_tell, with what would settle it.
Who uses it
Law firms
Line up the patta holder against the vendor in the draft sale deed before a title opinion.
Lenders
Flag a borrower whose name does not match the revenue record of the land offered as security.
Property platforms
Warn a buyer when the listed seller and the recorded holder look like different people.
Limits and good practice
- Advisory only. A match is not proof of identity; confirm with ID documents, the father's or spouse's name and the address on the record.
- A non-match is not proof of fraud. Registers carry initials, caste names, varied spellings and "and others" entries.
- A partial match usually means a family member, which needs the succession or settlement papers.
- The answer is written by a language model. Two calls on the same pair can word the reason differently.
Pricing
Each comparison costs 1 credit; nothing when the comparison could not run. Credits come from the same LandLens wallet as the chat and the minis. How billing works.
For thousands of calls a month, a portfolio sweep or a platform built on these registers, we quote a customised bulk price, with invoicing to match.
Questions
Is there an API to check if two names on land records are the same person?
Yes, this one. It compares two names as written and returns a verdict, a confidence and the reason.
How much does a name match cost?
1 credit per comparison. Nothing when the comparison could not run.
Does it handle Tamil names and initials?
Yes. Initials in any order, spelling variants, and a name in Tamil script against the same name in English are reconciled.
Does a match prove the seller owns the land?
No. It says the two names likely refer to one person. Identity is proven by documents, and title by the deeds and the EC.
Updated 22 Oct 2026 · OpenAPI description