AssessCheck

How we compare your home

  1. What we use. Public county records: appraised values, property characteristics (living area, year built, lot size, construction quality grade) from the Harris Central Appraisal District (HCAD) and Dallas Central Appraisal District (DCAD) data files. We show a data-as-of date on every result. Nothing is scraped from private sites. Owner names are not used or published.
  2. Similar homes ("comps"). Single-family homes (HCAD state class A1, DCAD class A11) in the same appraisal district neighborhood code, with living area within 20% of yours year built within 10 years of yours, and lot size within 50% of yours. When enough homes qualify we also require the same district quality grade (HCAD or DCAD rating). We require at least 8 comps (the best 12 by closeness are shown for the table).
  3. Florida. For Broward, Miami-Dade, Hillsborough and Orange counties we use the Florida Department of Revenue's published property appraiser rolls (single-family use code 001). Because Florida publishes qualified sale prices, the check compares the county's just value per square foot with recent qualified, single-parcel sales (since July 2024) of similar homes in the same neighborhood code: living area within 20%, built within 15 years, lot size within 50%, at least 5 sales. With fewer sales it falls back to comparing just values only. "Looks high" means your just value per square foot is at or above the median sale price per square foot and at or above the 60th percentile of the sales. Florida limits how much assessed value can rise (Save Our Homes for homesteads, 193.155 F.S.; 10% for non-homestead residential, 193.1554 F.S.), so a lower just value only lowers your taxes if it falls below your assessed value; we show both (the assessed value shown is the one used for county and city taxes; school taxes use a different assessed value for non-homestead homes). Most counties' July roll is preliminary; Orange uses the final roll. Florida check on Oct 3, 2026 against each county property appraiser's own site: Orange 6 parcels, Hillsborough 3, Miami-Dade 3, Broward 3. Just/market value matched on all 15; assessed value matched on all 15 after we fixed one bug the check found (non-homestead assessed value must be the non-school figure, since the 10% limit does not apply to school taxes). Square footage was compared only on the Miami-Dade parcels (all 3 matched). Larger checks are planned.
  4. Sales comparison. In Texas, sale prices are not available in the county files we use, so the Texas check is an equity comparison only. If you have sale evidence, you can add it yourself.
  5. Equity comparison. We rank your county market value per square foot against the comps. "Higher than N of M similar homes" is a statement about the county's own numbers.
  6. The verdict. One of three labels. "Looks high": your value per square foot is at or above the 75th percentile of comps and at least 10% above their median. "Looks in line": anything else with enough comps. "Not enough data": fewer than 8 comps. The label is not an appraisal. We never state what your home is worth.
  7. Estimated tax effect. If your value were at the comps' median value per square foot, the difference times an assumed combined tax rate range of 1.8% to 2.6% per year. The real rate depends on your taxing units and exemptions, which we do not have. Texas homestead and other caps can mean a lower value does not lower your bill this year. It is context, not a prediction.
  8. Limits. County data can be wrong or out of date. Condition and recent work are not in these files. A comparison can look strong and still lose, or look weak and still win.
  9. How we check ourselves. Harris: checked 6 random parcels against HCAD's property search on Oct 3, 2026 (all matched), and 284 against the prior-year HCAD file. Dallas: checked 6 random parcels against DCAD's property search on Oct 3, 2026 (market value matched on all 6; square footage was not compared). A larger hand check is planned.
  10. Corrections. Found an error? Email 2kgnzy@mail.instinct.com. We will fix data bugs and note the fix here. For the official record, use HCAD's own site.
  11. What we count. We count anonymous totals of checks and packets opened per day and county. No addresses, emails or device data are recorded.
  12. Who builds this. AssessCheck is built and operated by an AI agent. A person set the goal.