A foundation model of physical development

See how physical development changes next.

Try:

Explore the API Explore markets

Pick a market below. The chart and map stay linked, and selecting a map area redraws the forecast.

Brickell

Miami, FL

2027-2029 outlook

Median forecast (P50): +7% to $776K by 2029, with a modeled range from -16% to +35%. Model confidence is 73%.

Brickell

Interactive forecast map

Explore Map

Interactive forecast map - explore neighborhoods, tracts, and parcels

Explore Map

Tap a market above or use the map to inspect another area.

50M+

Parcel observations

1 to 5 Year

Horizons

~7%

Median error (1yr)

14M

Production requests each week

Production requests measure system activity, not people, customers, demand, or revenue.

A foundation model for physical development.

The live product begins with parcel and neighborhood forecasts that make time, geography, and uncertainty explicit.

Forecasts, not static records

Most built-environment data describes what exists today. Homecastr starts with how parcel and neighborhood values may change over the next one to five years.

See a live forecast

Range of outcomes

Lower, median, and upper estimates designed to express uncertainty explicitly rather than hide it behind a single estimate.

Lower range (P10)
Median forecast (P50)
Upper range (P90)

View scenarios

From parcels to places

Move from an individual parcel to its neighborhood and market context, with forecasts tied to a specific geography and horizon.

Explore the map

Comparable scenarios

Compare places, horizons, and modeled outcomes instead of relying on one static estimate or one set of historical comparables.

Read the methodology

Forecasts evaluated against outcomes.

Homecastr reports held-out results by geography and horizon, including about 7% one-year-ahead median error in current public evaluations.

Dated public forecasts are compared with realized outcomes. P10, P50, and P90 values report the modeled range for each forecast.

Read the methodology

Historical evaluation

Results reported across multiple market cycles.

Point-in-time inputs

Historical inputs are separated from the outcomes used for evaluation.

Percentile ranges

P10, P50, and P90 values show the modeled range.

Dated public records

Public records support training and evaluation for physical development.

The broader model is for teams shaping physical development

Planners and public agencies

Compare independent scenarios with prior public forecasts and measured outcomes.

Developers and investors

Test long-range theses across parcels, neighborhoods, and market conditions.

Utilities and infrastructure teams

Explore how demand, capacity, hazards, and development interact over time.

Model and data teams

Access forecast data through REST, MCP, and bulk interfaces.

API Access

The live forecast layer, via API.

Access parcel-level and neighborhood-level forecasts programmatically. JSON responses, API key auth, sub-second latency.

GET /api/v1/forecast

{
  "address": "1000 Brickell Ave, Miami, FL",
  "current_value": 295000,
  "forecasts": {
    "p10": 268000,
    "p50": 345000,
    "p90": 425000
  },
  "horizon_years": 5
}

Get Free API Key API Documentation