How Vurel works, plainly

Every night, Vurel reads publicly posted property listings from 9 major Thai portals: LivingInsider, Property Hub, Thailand-Property, DDproperty, FazWaz, Dot Property, Kaidee, Teedin108, and KaiBaanThai. Across those sources we track roughly 1.4M listings covering 47+ Bangkok zones, for both sale and rent, in English and Thai.

We collect only what a portal already shows to anyone who visits it: the project, the unit specification, the zone, the asking price, and any contact the seller chose to publish. We do not scrape private accounts, gated data, or anything behind a login.

The same unit is often posted to several portals at once, sometimes by different agents at different prices. Vurel keeps each of those as a separate listing tied to the portal it appeared on. We do not merge or de-duplicate the same property across portals.

This is a deliberate choice. Cross-portal de-duplication for Thai listings is unreliable, and a wrong merge hides real price differences. So a count of Vurel listings is a count of portal postings, not a count of distinct physical units. We would rather show you nine honest postings than one guessed-at merge.

For each listing with a usable price and floor area, we compute price per square metre as the asking price divided by the unit area in square metres. To describe a zone, we take the median of those per-sqm values, not the average. The median is far more robust to a handful of mispriced or luxury outliers than a mean.

Before computing a zone figure, we trim extreme outliers, listings whose per-sqm value sits far outside the bulk of the zone, since these are usually data-entry errors, mislabelled land, or non-comparable stock. Where a zone has too few listings to be meaningful, we show the sample size rather than a confident-looking number built on three data points.

Each night we re-check the listings we have seen before. When a listing's asking price changes, we record the new number alongside the date. A price cut is simply a night where a listing's asking price dropped below its previous recorded asking price.

This is the memory the portals do not keep. A portal shows you one number today. Vurel shows you every asking price that listing has carried since the first night we recorded it, which for most listings begins around March 2026. We do not claim years of backfilled history we never collected.

Vurel runs on a nightly cycle. Each morning reflects the most recent completed overnight read, not a live ticker. A price changed at noon will appear after that night's run, not the instant it happens. We say "last snapshot: tonight," and we mean it literally.

Roughly 2 in 3 listings (~64%) carry a contact the seller published, a phone number, a LINE ID, or an email. The rest do not, and we do not invent one. Many published contacts are agency desks rather than the direct owner. We show what was posted, masked in previews, and never imply we hold every owner's number.

This is YMYL territory: people make money decisions on it. So the scope has to be stated plainly.

• Is, asking prices

  Every price is the price a listing asked, taken straight from a public portal.

• Is not, sold prices

  We never show what a unit sold for. There is no link to the Land Department registry or any transaction record.

• Is, nightly snapshots

  Each figure reflects the latest completed overnight read of the portals.

• Is not, real-time or an appraisal

  Vurel is market context, not a live feed and not a certified valuation. Use it to inform a decision, not to replace professional advice.

• Is, per-portal postings

  Counts are of listings on the portals they appeared on, deliberately not merged across sites.

• Is not, distressed or auction data

  No NPA, foreclosure, or auction stock. Only ordinary public for-sale and for-rent listings.

We are happy to explain any figure or show our working. If something here is unclear, or you spot a number that looks wrong, write to us at hello@vurel.io.