Our Methodology
Transparency is one of our core principles. Here's exactly how we calculate your home's value.
Data Sources
MLS Listings
Active and historical listing data including list prices, days on market, and price changes.
Public Records
County tax assessments, deed transfers, and property characteristics from assessor databases.
Permit Data
Building permits reveal renovations, additions, and improvements that affect value.
Market Analytics
Real-time market conditions, inventory levels, and price trends at the neighborhood level.
School Ratings
School district boundaries and ratings from GreatSchools, which significantly impact home values.
Neighborhood Data
Demographics, crime statistics, walkability scores, and amenity proximity.
The AI Model
Our valuation model is a gradient-boosted ensemble trained on over 50 million residential sales transactions from the past 10 years. The model learns complex relationships between property features, location factors, and sale prices.
Key model characteristics:
- Trained on 50M+ transactions across all 50 states
- 500+ features including property, location, and market variables
- Retrained weekly to capture current market conditions
- Separate sub-models for different property types and price tiers
- Confidence intervals based on comparable sale availability
We also use a large language model to generate plain-English explanations of the valuation factors, making the results accessible to non-experts.
Accuracy & Limitations
Where We're Most Accurate
- Major metropolitan areas
- Standard single-family homes
- Areas with high transaction volume
- Properties with recent comparable sales
Where Accuracy May Vary
- Rural areas with few sales
- Unique or custom-built homes
- Recently renovated properties
- Multi-family and commercial
Our valuations are estimates and should not be used as a substitute for a professional appraisal. For legal, mortgage, or investment purposes, consult a licensed appraiser.