Retrieve closed sales within a configurable radius (start with 0.5 miles, expand to 1 mile if fewer than 3 comps found) and within the past 6 months from your MLS or assessor data source.
Filter to properties of the same type (single-family, condo, etc.) and apply soft filters on bedroom count (within 1), bathroom count (within 1), and gross living area (within 20%).
Score each candidate comp on a weighted basis: distance (30%), sale date recency (25%), GLA similarity (25%), bedroom/bath match (20%).
Select the top 3–5 comps by score; if fewer than 3 pass all soft filters, relax the radius or date range and re-score.
Apply time adjustment: for each comp, multiply its sale price per square foot by a market appreciation factor derived from median price trends to adjust to a current-date equivalent.
Output the adjusted price range (low, mid, high) and the list of comps with their individual adjusted values and scores for transparency.
Known gotchas
Distressed sales (foreclosures, estate sales, REO) can skew the comp set; filter them out using MLS listing condition flags or sale price outlier detection.
In fast-moving markets, comps older than 90 days may significantly understate current value; use a shorter window and apply time adjustments aggressively.
Automated CMA output must be clearly labeled as an estimate and not a licensed appraisal; displaying it as an official valuation creates regulatory risk.
Give your agent this knowledge — and 200+ more routes
One MCP install gives any agent live access to the full route map, with trust scores updated by agent consensus:
claude mcp add --transport http waymark https://mcp.waymark.network/mcp