“Quick answerPlan on two or three property data APIs, because no single provider covers every field. Homesage.ai returns estimated values, property condition scores and estimated returns, with insights on over 155M US properties. It is the only one of the 7 that scores a home’s condition from its photos and returns an estimated renovation cost for US properties. ATTOM supplies bulk records, Cotality covers mortgage and risk data, and Zillow Group licenses listing data to approved partners. HUD Fair Market Rents and the Census ACS are free public sources for rent benchmarks and demographics, and Walk Score adds walkability.
Postman’s 2025 State of the API Report found that 82% of organizations have adopted some level of an API-first approach. The 2024 edition reported 74%. In real estate the dependency runs deeper, because almost none of the data originates with you.
The useful question is which two or three APIs you combine, and in what order.
Why the data layer decides what your product can do
Fortune Business Insights projects the global PropTech market at $44.59 billion in 2026, with integrated platform and software solutions taking 65.24% of that share. Behind nearly all of it, a property data API is doing the work.
Four things separate a usable API from a good one: coverage depth, update frequency, schema consistency and developer experience. A fifth, what the license lets you store, constrains all four.
Why one API is never enough
The argument for layering is usually made with architecture diagrams. It is easier to show with a market.
We read one city through three signals on the same day: active detached listings in Columbus, Ohio, on 17 October 2026. The first thing to measure is how many listings carry the data to be judged at all.
| Columbus population | Listings | Share of the market |
|---|---|---|
| All active detached listings | 1,270 | 100% |
| Carrying a condition score | 1,151 | 90.6% |
| Carrying a return grade | 1,270 | 100% |
| Carrying a cap rate | 87 | 6.9% |
| Carrying all three | 84 | 6.6% |
Our property database, Columbus active detached listings, 17 October 2026. A cap rate is only computed where a verified rent exists, which is what holds that row to 6.9%. Counts move daily.
Two of the three signals cover almost the whole market. The third covers 87 listings. Any comparison across all three has to run on the 84 that carry all of them. A listing with no cap rate has not failed a yield test. It is simply unevaluable.
On those 84, the signals still disagree. 25 are outdated or worse, 8 clear a 6% cap rate, and 1 carries a high return grade. Each signal gives a different answer on the only listings where all three can be asked.
Count the listings that carry the data before you count the ones that pass a filter.
The 7 best real estate data APIs
The 7 at a glance (checked 22 October 2026)
| Provider | What it returns | How you get access | Best for |
|---|---|---|---|
| Homesage.ai | Estimated values, comps, rent estimates, property condition scores, estimated renovation costs, investment scores | API key from the developer portal after you create an account | Investment analysis and deal tools |
| ATTOM | Property, tax and sales records for 160 million+ US properties, an AVM, a rental AVM and comps | 30-day trial key | Underwriting, insurance and enterprise records |
| Cotality | Property characteristics, valuation, climate risk scores, market trends. Trestle for MLS feeds | Account on its developer portal. Trestle once each MLS approves | Mortgage and risk |
| Zillow Group | Zestimates, public data, rentals and listings, by API family | Public or partner access. The MLS listings API is invite only | Consumer search displays |
| HUD Fair Market Rents | Rent benchmarks for every county, by bedroom count | HUD USER account and a token | A rent fallback |
| Walk Score | Walk Score, Transit Score and Bike Score | API key, with a free tier and a daily quota | Location context |
| US Census ACS | Population, income, employment and housing tenure | Up to 500 queries a day without a key | Neighborhood demographics |
1. Homesage.ai. Best for AI-derived property intelligence
Our real estate API returns analysis as well as records, with insights on over 155M US properties. The models draw on assessor records, transaction history and property photos. It is the only provider on this list that scores a home’s condition from its photos and returns an estimated renovation cost for US properties. ATTOM and Cotality return the county assessor’s condition code instead.
Six groups of endpoints matter most to investment products. Property condition scores a home from its property photos on a five-level scale. Estimated renovation cost breaks the work down room by room. Investment potential scores the return.
Seller Motivation Score estimates how much room a seller has to negotiate. The rental endpoints estimate long-term and short-term income. The valuation and data endpoints return estimated values, comps and structured records.
Every endpoint has a reference page with an example request and a sample response. The developer playground then runs a live call with your own key before you write any integration code.
Best for: investment platforms, lending dashboards and anything that sells a decision rather than a record. Our IT developer tools cover the integration path.
2. ATTOM. Best for enterprise coverage and normalization
ATTOM states coverage of more than 160 million US properties. Its AVM combines county assessor records, county recorder records and MLS data, and its Enterprise Data Management Program normalizes and validates across those sources.
That normalization is the differentiator. If you need one schema to behave the same way in every county, ATTOM is built for it.
Best for: mortgage underwriting tools, insurance applications and anything financial-grade. Our comparison of when to use each real estate API goes further.
3. Cotality. Best for mortgage and risk data
Cotality is the company that traded as CoreLogic until its rebrand on 24 March 2025. If your vendor list still says CoreLogic, that is the same business.
Its developer portal lists property characteristics, valuation, climate risk scores and market trends. Trestle, its MLS data distribution product, still carries the Trestle name. Cotality says it serves 105 MLSs and more than 1,250 technology providers. The 5 predictive analytics tools lenders use shows the same stack inside a lending desk.
Best for: mortgage platforms, title companies and risk modeling.
4. Zillow. Best for consumer listing context
Zillow’s 2025 annual report describes a living database of approximately 173 million US homes. Its Zestimate page lists a median error of 1.77% for homes on the market and 7.20% for off-market homes.
Zillow Group offers some APIs to the public and others to approved partners, and its MLS listings API is invite only.
Read the terms before you design around it. Zillow’s developer terms bar retaining copies of the data and allow no more than 20 properties on screen for a user at a time. They also bar using the data to provide a service for other businesses.
Best for: consumer-facing search and baseline displays, inside the license you are granted.
5. HUD Fair Market Rents. Best for a free rent benchmark
The US Department of Housing and Urban Development publishes Fair Market Rents for every metro and non-metro county, by bedroom count, updated annually. FY2027 rents took effect on 1 October 2026.
The API is free and returns clean JSON. It needs a free HUD USER account and a bearer token, which takes a minute. This is the fallback for the gap the Columbus count exposed, because a published benchmark covers the addresses where no verified rent exists.
Best for: rental platforms, landlord dashboards and any rental math that must not return null.
6. Walk Score. Best for location context
Walk Score returns walkability, transit and bike scores for addresses in the United States and Canada. It is lightweight enough to drop into a property card without a second thought.
There is a free tier with a daily quota and paid tiers above it. Check the status field in every response. A call over the quota still returns HTTP 200, with status 41 in the body, so an unchecked quota shows up as a blank field.
Best for: buyer-facing search and any scoring model where location quality carries weight.
7. US Census ACS. Best for neighborhood demographics
The Census Bureau’s American Community Survey API is free and authoritative for population, income, employment and housing tenure. The 2020 to 2024 five-year release landed on 29 January 2026.
It allows 500 queries per IP per day without a key, and a free key lifts that. Treat it as a background context layer rather than a live feed, because it updates annually.
Best for: market research tools and neighborhood comparison features.
Which combination fits which platform
| Platform type | Primary API | Supplementary APIs |
|---|---|---|
| AI investment analysis tool | Homesage.ai | ATTOM, Walk Score |
| Mortgage underwriting platform | Cotality | ATTOM, Census ACS |
| Rental investment platform | Homesage.ai | HUD Fair Market Rents, Walk Score |
| Buyer search portal | Zillow | Walk Score, Census ACS |
| Hard money lending dashboard | Homesage.ai | Cotality, ATTOM |
| Market analytics product | ATTOM | Cotality, Census ACS |
Published capabilities and terms, checked 17 October 2026.
How to layer the stack
Build in four layers. Core records first, because everything joins to the parcel. Then the specialist analytics your product actually sells. Then location context. Then the free public fallbacks that keep a field from coming back empty.
Before you sign anything, our guide to real estate API implementation practices covers rate limits, caching and fallback handling.
Key takeaways
- No single provider covers every field, so plan on two or three property data APIs in production.
- Layer them in build order: core records, specialist analytics, location context, then free public fallbacks.
- In Columbus on 17 October 2026, only 84 of 1,270 listings (6.6%) carried a condition score, a return grade and a cap rate at once.
- Free public benchmarks such as HUD Fair Market Rents cover the fields a paid feed leaves empty.
- Read the license before the docs, because what you may store constrains the architecture more than any rate limit.
Conclusion
The API you pick shapes your feature ceiling for years. It is the least reversible decision in the build.
Condition scores, estimated renovation costs and estimated returns are fields your users will not assemble themselves. An API that returns them saves you building those models.
Pick the core, pick the analytics, and then plan honestly for the fields that come back empty.
Create a Homesage.ai account and call the analytics endpoints from your own stack.
Frequently asked questions
What is a property data API?
It is an interface that returns structured data about a property, such as ownership, transaction history, valuation and physical characteristics. Your application queries it rather than maintaining the records itself.
Which property data API has the best US residential coverage?
ATTOM publishes coverage of more than 160 million US properties, and Zillow reports a living database of approximately 173 million homes. Coverage counts measure different things, so compare the fields you actually need.
How many APIs does a PropTech platform typically need?
Two or three in production. One for core records, one for the analytics your product sells, and one for location or demographic context.
What should developers check before committing to a provider?
Update frequency, schema consistency across counties, a way to test before you pay, and the license terms on storage. Test one known property through every candidate before you sign.
Sources
- Postman, 2025 State of the API Report, seventh annual edition, published 8 October 2025.
- Fortune Business Insights, PropTech market size, page updated 28 September 2026.
- Cotality, CoreLogic global rebrand announcement, published 24 March 2025.
- ATTOM, property data coverage, checked 17 October 2026.
- Our property database, Columbus active detached listings, 17 October 2026.
Estimated values, estimated ARV, estimated renovation costs and estimated return scores are model-derived estimates, not investment advice. Property data as of 17 October 2026.
