Property records and parcel MCPs
A property records MCP lets an AI look up a property's owner, tax assessment, sales, mortgages and parcel boundaries. Official ones come from ATTOM, Cotality and Regrid.
[Top button: Records are step one. See steps two through five. → demo]
What's in a property record
Data companies collect records from county assessors, recorders and courts, then standardize them nationally.
- Ownership: owner name(s), mailing address, owner-occupied or not, and whether the owner is a person, trust or LLC.
- Tax: assessed value, annual tax, and whether taxes are paid.
- Transfers: past sales, prices, dates and deed types.
- Mortgage: recorded loans, lenders, amounts and dates, so you can estimate equity.
- Property: beds, baths, square feet, year built, lot size, land use.
- Parcel: the legal boundary, APN (parcel number) and zoning.
ATTOM
ATTOM is one of the largest US property data companies, covering ownership, tax, deeds, mortgages, AVMs and neighborhood data. Its official MCP server covers property details, valuations, sales history and comps.
Full detail on our ATTOM MCP page.
Cotality
Cotality, formerly CoreLogic, has an official MCP server for property, market, risk and valuation data. It's for licensed customers and is in Claude's connector directory.
Regrid
Regrid covers US parcel boundaries, owners and land data. Its official MCP server uses your API token and bills per parcel record.
More on our Regrid review.
County and parcel servers
Two community projects go straight to the source:
- UrbanKit Atlas (leoyrh/mcp-atlas) finds verified county parcel map endpoints (ArcGIS) and builds owner or APN queries.
- cornerstone-mcp (jsconiers/cornerstone-mcp) is a forkable "real estate intelligence" server with about 156 tools across listings, valuation and public data.
One address, every server
One property, looked up through each records server we could access.
| Field | ATTOM | Cotality | Regrid | Goliath |
|---|---|---|---|---|
| Owner name | [fill] | [fill] | [fill] | [fill] |
| Mailing address | [fill] | [fill] | [fill] | [fill] |
| Assessed value | [fill] | [fill] | [fill] | [fill] |
| Last sale | [fill] | [fill] | [fill] | [fill] |
| Mortgage | [fill] | [fill] | [fill] | [fill] |
| Parcel boundary | [fill] | [fill] | [fill] | [fill] |
| Distress filings | [fill] | [fill] | [fill] | [fill] |
Pricing models: how records servers charge
| Model | Who | What it means for you |
|---|---|---|
| Subscription plus quote | ATTOM, Cotality | A sales conversation and a monthly minimum. Built for data teams. |
| Per record | Regrid | You pay for what you pull. Cheap to try, harder to predict at volume. |
| Included in a software plan | Goliath | Records come with the CRM plan; owner lookups spend the plan's export credits. |
Watch usage when the AI does the querying, because chained lookups pull records fast. Our cost page adds up a full stack of separate tools.
From record to owner to conversation
A record tells you the owner. To do a deal you still need to:
- Pick the right properties. Usually by a reason to sell: a pre-foreclosure, a probate case, back taxes.
- Reach the owner. That takes a phone number or email, from skip tracing.
- Track the conversation: calls, texts, notes and follow-ups, in a CRM.
- Run the deal: comps, an offer, a contract.
Questions
Can I get property records into ChatGPT for free? Partly: community servers that query county parcel maps are free, but coverage varies by county.
Do property records include phone numbers? No, only the owner's name and mailing address.
What's the difference between property records and MLS data? Records cover every property from public sources, while MLS data covers listed properties, from agents.
Records are step one. See steps two through five. Watch Goliath go from filing to owner to call list from one AI conversation, in a 20-minute demo.