LATAM real estate intelligence
REAL NOW · Real estate intelligence built for LATAM
We build the analytics and decision tools that investors, developers, brokers and banks across Latin America use to operate with rigor — not intuition. Daily refresh, multi-country coverage, institutional-grade quality controls.
7
LATAM countries
Daily
Pipeline refresh
5
AI personas
ES · EN
Bilingual platform
What we do
A complete intelligence stack for operators, not a listing portal.
01
Zone-level analytics.
Asking-price percentiles, $/m² medians, days-on-market, estimated sales, repricing radar and momentum — computed per zone, property type and bedroom segment.
02
Neighborhood quality scoring.
Walkability, services, schools, transit, healthcare and recreation — scored asymptotically so dense areas differentiate honestly and suburbs are not flattered.
03
Institutional AI briefings.
Five expert lenses — Investor, Developer, Seller, Bank/Appraiser and Broker — turn raw data into a verdict + confidence + rationale, in Spanish or English.
04
Verified broker inventory.
A public marketplace built from broker-uploaded listings only, after KYC and per-listing admin approval. No anonymous postings, no scraped resale leakage.
05
Project demographics.
Aggregate signals on new construction across LATAM: what is coming, delivery years, typical sizes, price tiers, developer concentration and amenity adoption.
06
Risk + civic event layer.
Adverse events (crime, accidents, hazards, protests) classified per country and department, feeding zone safety scoring and address-level alerts.
Data quality and process rigor
How we earn the right to call this institutional.
01Daily refresh pipeline
A 27-step orchestrator runs every day to refresh listings, news, hazards, macro indicators and POIs. Each step is tracked individually with success/failure telemetry.
02Silent over wrong
When a parse, a price or a classification is uncertain, the field stays NULL. Confidently wrong is worse than honestly blank — every analytic we publish inherits this discipline.
03Three-layer observability
Sanity bounds at the parser, anomaly detection on the merged dataset, and a daily quality report posted to internal Slack — anything off-spec is caught within 24 hours.
04Source reliability tiering
Public information is classified by whether it removes listings that leave the market. Movement metrics — days-on-market, estimated sales, repricing — only count signals from sources that do it faithfully.
05Dual cross-validation
Geocoding runs through two independent providers and disagreements are flagged before anything reaches the public surface. The same principle applies to price parsing and classification.
06Curation with audit trail
Admin operators can flag, edit or remove individual rows; every change is timestamped and stored in an audit log. Flagged items auto-exclude from movement views — analytics never quote bad data.
Estimated sales: how we calculate them
No public registry of closing prices exists in our markets, so we never present sales as registered closings. What we show is an estimate, with its method declared.
01
What we count.
Sale listings that left the published supply in the last 30 days. Rentals are never mixed in.
02
How much each exit weighs.
If the source marks the listing sold, it counts as 1 sale. If we confirmed the listing is gone (its page no longer exists or it left the full index of the source), it counts as 0.5: we estimate one chance in two that it sold rather than being withdrawn or rented. If it simply stopped appearing on a portal that removes listings, it counts 0.5 times the measured probability that it really left that portal, discounting the share republished under another listing. Disappearing from portals that we read by sample counts 0.
03
What is measured and what is assumed.
Whether a listing really left is measured: we track how many come back, and we re-measure every month. Whether it sold or was withdrawn cannot be measured with public listings; the 0.5 is a stated assumption until brokers report real closings.
04
Prices and time.
Prices are the last list price we saw, not the closing price. Days on market is how long active listings have been published, not how long a property takes to sell.
05
Exits are not sales.
Exit rates count every listing that left the published supply — sold, rented or withdrawn — so we show them as exits, never as sales.
Our stance
Locked positions — these never change without a public note.
Analytics first, never a marketplace.
We do not publish individual scraped listings to the public. Aggregates only. The public inventory comes from KYC-verified brokers who chose to be there.
Public sources stay anonymous.
We never name specific portals or vendors publicly — even when our pipeline reads them. Source attribution lives only in operator-side admin tools.
LATAM-first payments rail.
Subscriptions and one-time purchases settle through Stripe — secure card payments in USD. Built to scale across our LATAM markets.
Bilingual by default.
Every page, every briefing, every email digest is available in Spanish or English. AI briefings live in separate caches per language so you never get translated leftovers.
Coverage
8 LATAM markets active today, with the framework to expand without breaking analytics consistency.
Built for
Different operators, different questions, the same underlying truth set.
Investors
Yield, cap-rate proxy, repricing radar, peer-zone comparison.
Developers
Land prices, estimated sales, segment whitespace, demand mix.
Banks and appraisers
Comparables, volatility, days-on-market, LTV signals.
Brokers
Stale inventory leads, motivated-seller signals, verified portfolio surface.
Buyers
Address analysis, zone fit, affordability calculator, mortgage scenarios.
Who operates REAL NOW
REAL NOW is operated and billed by ARMKU LLC, a data company based in Houston, Texas (USA).
We publish market data and analysis. We are not a real estate broker, we do not intermediate sales and we never hold third-party money.
Contact: hola@realnow.com