ChatGPT App Walkthrough: Zillow

Zillow’s ChatGPT App might be one of the most complex integrations launched so far. It turns ChatGPT into a lightweight real estate assistant—able to browse properties across the U.S., compare different neighborhoods, estimate how much home you can afford, and even calculate estimated mortgage and interest payments.

We tried it out to document the full breadth of capabilities and see how it might be helpful to people searching for new apartments, homes, or neighorhoods.

---

Getting Started: What Zillow’s ChatGPT App Does

Before you connect anything, ChatGPT can suggest the Zillow App when you ask about home prices, mortgages, or where to live.

Much like on its website, Zillow doesn’t require any login to access the app. However, ChatGPT does require you to ‘connect’ the app, which is one step and doesn’t require any authenticaiton:

Once connected, you can ask things like:

• “What kind of house can I afford in Austin if I make $120k?”
• “Show me neighborhoods near Seattle with good schools.”
• “What’s the monthly payment on a $600k home with 10% down?”

Zillow can handle all of these, in many cases with an interactive map or detailed calculations included. Interestingly, for ‘show me neighborhoods’ type searches, ChatGPT specifies the actual neighborhoods, and just tells Zillow why they were selected, rather than the other way around:

Zillow just returns the corresponding data on the neighborhoods and their housing options.

---

The Four Core Tools

1. Zillow Property Search Tool

This is the main property search engine. It can find listings for sale or rent anywhere in the U.S. with filters for:

• Price range
• Bedrooms and bathrooms
• Property type (house, condo, apartment)
• Commute time or nearby location
• Monthly cost

You can combine these in natural language:

• “Show me two-bedroom apartments under $3,000 near Golden Gate Park.”
• “Homes within 15 minutes of downtown Kansas City.”
• “Find ADA-compliant rentals in Austin.”

It returns real listings, complete with photos, price, location, and links back to Zillow.

Interestingly, Zillow shows you what filters are being used, but requires you to use natural language to update them:

However, within a given property listing, there’s quite a bit of interacitvity that users can access directly within the Zillow ChatGPT App:

---

2. Calculate Home Affordability Tool

This tool runs Zillow’s BuyAbility℠ model from Zillow Home Loans. It estimates how much house you can afford based on your income, down payment, credit score, monthly debts, and location.

It returns:

• Target and maximum home price estimates
• A detailed monthly payment breakdown (principal, interest, taxes, insurance, HOA)
• Estimated interest rate and APR from Zillow Home Loans

It works even if you only give partial information, using defaults for missing values. All results are labeled clearly as estimates, not financial advice.

After you’ve shared this information, listings also inform you of whether or not they’re within your ‘buyability’ as defined by the tool, which is a nice touch:

---

3. Interest Rate Mortgage Payment Simulator Tool

If affordability is about what you can buy, this tool focuses on how much it costs. It simulates monthly mortgage payments using real-time interest rates from Zillow Home Loans LLC (NMLS ID#: 10287).

You can ask:

• “What are today’s mortgage rates in Illinois?”
• “How much will my mortgage be on a $400k house with 10% down?”
• “How does my credit score affect my monthly payment?”

Results include rate comparisons, total payment estimates, and side-by-side scenarios that help users understand how small rate changes affect affordability.

As a push toward conversion, Zillow also includes a CTA within their ChatGPT app to apply for mortgage pre-approval. However, this experience doesn’t capture the info you’ve shared with ChatGPT - you end up back at the landing page and have to tell Zillow again about your geography, income, and more, info you likely already shared within the tool:

---

4. Zillow City/Neighborhood Real Estate Information

This tool is all about context, helping users who may not know where they want to live to explore U.S. cities and neighborhoods.

It returns descriptions and for-sale/rental property counts for up to 20 regions at once, with short summaries like:

“Capitol Hill is a residential neighborhood in Seattle known for walkable streets, cafes, and access to parks.”

It’s built to handle queries like:

• “Best neighborhoods in Austin for young professionals”
• “Areas near Denver with good hiking access”
• “Places in Boston close to public transit”

Zillow enforces strict Fair Housing Act compliance: no demographic targeting, exclusionary filters, or steering based on protected characteristics. The system is designed to provide neutral, amenity-based recommendations.

As might be expected, the context shared by the Zillow ChatGPT App is heavily grounded in apartment/home inventory available for sale or rent, and approximate cost of living/purchasing power by geography. However, for those topics, ChatGPT now can provide helpful summaries using Zillow’s actual data:

---

What’s Not Supported

The Zillow ChatGPT App only covers the U.S. and doesn’t provide financial advice or lending commitments.
It doesn’t support demographic, racial, or age-based searches due to Fair Housing laws, and all outputs are explicitly neutral.

You also can’t start a loan application, submit offers, or schedule tours directly from ChatGPT yet; it’s informational only.

---

Where It Could Go Next

As ChatGPT’s memory and API context improve, Zillow could personalize results based on prior interactions. Imagine asking, “Can you find houses I could actually afford in the neighborhoods we discussed earlier?” and getting a merged output from all four tools.

For now, it’s a clean, transparent, and compliance-safe foundation for how real estate apps might work inside conversational AI.

---

Final Take

Zillow’s ChatGPT App doesn’t try to replace the Zillow site. It gives users a way to ask housing questions conversationally and get grounded, data-backed answers right inside ChatGPT.

It’s not flashy—it’s structured, compliant, and reliable. That might make it one of the most practical ChatGPT Apps launched so far.

---

Appendix: Full Zillow ChatGPT Actions Metadata (as of 11/16/25)

zillow_property_search

Zillow Property Search searches for U.S. real estate properties (for sale or for rent). Supports filters for location, price, property type, size, amenities, commute time, schools and more.

What it returns:

• A set of matching property listings
• Includes location, photos, and details based on user-specified filters

Inputs:

• propertySearchRegionFiltersRequest (optional): Defines the geographic search area(s) and context
• propertyFiltersRequest (optional): Only include filters the user specifies (omit others; do not send placeholder unknown)
• displayLimit (optional): Maximum number of properties to return (default: 100; only set if user requests)

Key Rules & Constraints:

• U.S. properties only
• Max 5 regions per request
• Only set filters explicitly provided by the user
• Do not default any enums to unknown
• Use structured filters, not free text
• DO NOT SEND Sort Selection (livingArea etc.) unless user EXPLICITLY asks to sort
• Send only one of: school, schoolDistrict, propertySearchRegions, or customPolygonWkt
• Do not set sortSelection or displayLimit unless user explicitly requests it
• Always use nearby sorting along with locationForNearbySort
• Requests need to adhere to the input schema
• Re-call tool if user modifies search criteria
• Safe defaults are applied if no filters are provided

Fair Housing Compliance:

• Must comply with U.S. Fair Housing Act and applicable state and local laws, including those that prohibit source of income discrimination in rental transactions
• Protected Classes: Race, Color, Religion, Sex (including Sexual Orientation and Gender Identity), National Origin, Familial Status, Disability
• Also treat as protected where applicable: Source of Income, Military/Veteran Status, Age (55+ exemptions may apply), Criminal History
• Prohibited: Any preference, steering, or exclusion based on protected characteristics

Special Features:

Nearby Searches: Use sortSelection: "nearby" along with locationForNearbySort (lat/long anchor) for queries such as “Homes near Coit Tower, San Francisco”

Commute Filtering: Use CommuteTimeFilter for commute-based queries (e.g., “Homes within 15 minutes of Kansas City”). Requires destination coordinates, commute time, mode, and timeOfDay

Monthly Cost Filtering: Use monthlyCostFilter for monthly cost-based queries. monthlyCost field is REQUIRED. Defaults will be applied for everything else

School Search: Use schoolName for searching by school (e.g., “Comstock Elementary School, McKinney, Tx”). IMPORTANT: Use School Name with city & state

Display Limit: Never set displayLimit in requests. The server will handle defaults. Ignore displayLimit completely, even if you want fewer results, unless the user explicitly says ‘limit results’ or gives a number

Example Minimal Request:

{
  "request": {
    "propertyFiltersRequest": {
      "bedrooms": {
        "min": 2
      }
    },
    "propertySearchRegionFiltersRequest": {
      "propertySearchRegions": [
        {
          "description": "Seattle, WA is a major Pacific Northwest city ....",
          "label": "Vibrant Tech Hub",
          "name": "Seattle, WA",
          "regionDisplayName": "Seattle"
        }
      ],
      "userSpecifiedPivotArea": "Seattle, WA"
    }
  }
}

Metadata:

• Output template: ui://widget/zillow-property-search.html
• Invoking message: “Working with Zillow to find properties”
• Invoked message: “Loaded properties with Zillow”

---

calculateHomeAffordability

Calculates how much house a user can afford. Tool works with no inputs and partial inputs. Affordability depends on gross annual income, down payment, credit score, monthly debts, and location (within the United States of America). Returns the maximum affordable home price, estimated monthly payment, the Zillow Home Loans interest rate with APR, and a detailed breakdown including principal and interest, property tax, homeowners insurance, PMI, and HOA dues. Powered by Zillow Home Loans LLC (NMLS ID#: 10287) BuyAbility℠ tool.

Disclaimer: All results are estimates and not financial advice. This tool returns an illustrative mortgage rate estimate based on limited inputs and assumptions. Actual rates and APRs depend on underwriting, credit approval, and may change without notice.

Inputs (all optional):

• annual_income: Total yearly household income before taxes
• down_payment: Amount available for down payment
• credit_score: Credit score (affects interest rate)
• monthly_debts: Total monthly debt obligations
• location: City, state, or ZIP code
• monthly_mortgage_budget: max amount the user wants their mortgage payment to be

Outputs:

• Target Estimate, the maximum affordable home price at a 36% DTI or with user’s monthly budget
• Max Estimate, the home price the user may be eligible for
• Home price estimates are based on inputs provided and/or assumed
• Loan Information (Loan Type, Loan Interest Rate, Loan APR, Points Paid Down for this Interest Rate, Dollar Amount required to purchase the Points Paid)
• Interest rates used are provided by Zillow Home Loans LLC (NMLS ID#: 10287)
• Estimated monthly payment with detailed breakdown: Principal & Interest, Property Tax, Homeowners Insurance, PMI (if applicable), HOA Dues (if applicable)
• Annual income used, Down payment used, Monthly debts used, Credit score used, Location used

Important Notes:

• Missing inputs will use sensible defaults
• Results are estimates only - not financial advice or lending commitment
• Only supports U.S. locations
• Source: Powered by Zillow Home Loans LLC (NMLS ID#: 10287) BuyAbility℠ tool

Fair Housing Compliance: This tool must comply with the U.S. Fair Housing Act and related laws.

✅ APPROPRIATE Uses (affordability-focused):

• “What can I afford with $80k income and $20k down payment?”
• “Calculate affordability in Seattle, WA with my $5k monthly budget”
• “Show me what price range I qualify for with a 720 credit score”
• “What’s my buying power in White Plains, NY?” (location name, not discrimination)

❌ INAPPROPRIATE Uses (discriminatory):

• “What can my Hispanic clients afford?” (ethnicity-based steering)
• “Show affordability in areas without Section 8” (source of income discrimination)

Metadata:

• Output template: ui://widget/zillow-buyability.html
• Invoking message: “Working with Zillow Home Loans (NMLS ID#: 10287) to calculate affordability”
• Invoked message: “Loaded affordability calculator with Zillow Home Loans (NMLS ID#: 10287)”
• Widget CSP

---

interestRateMortgagePaymentSimulator

This tool helps users simulate their monthly mortgage payment based on real-time interest rates from Zillow Home Loans LLC (NMLS ID#: 10287). Unlike the Home Affordability Calculator, which estimates the maximum home price a user can afford based on their income and debt, this tool is focused on:

• Exploring daily mortgage rates in general or by location
• Calculating estimated monthly payments based on a specific home price, down payment, and credit profile
• Helping users understand how changes in interest rates, credit score, location, or down payment size affect their monthly mortgage costs

Inputs (all optional):

• Home price
• Down payment
• Credit score
• Property location (city/state or just state)

Use this tool when users ask:

• “What are today’s mortgage rates?”
• “What are today’s mortgage rates in Illinois?”
• “How much will my mortgage be if I buy a $400k house with 10% down?”
• “How does my credit score affect my monthly mortgage?”
• “Monthly payment for a home in Austin with 7% interest”

Key Capabilities:

• National Daily Rate Access – Retrieve up-to-date mortgage interest rates for the U.S. overall
• State-Specific Daily Rate Access – Request current rates for a particular U.S. state
• Payment Estimation – Calculate monthly payments based on loan amount, term, and rate
• Scenario Simulation – Show how payments shift as interest rates rise or fall
• Neutral Presentation – Provide information in a consistent, non-discriminatory way

Disclaimers:

Informational Use Only: This tool returns an illustrative mortgage rate estimate based on limited inputs and assumptions. It does not constitute financial, legal, or lending advice.

Variability of Results: Actual rates and APRs depend on underwriting, credit approval, and may change without notice. This is not a commitment to lend.

Fair Housing Compliance: This tool is designed to comply with Fair Housing laws. It does not consider or provide information based on personal characteristics protected by law (such as race, ethnicity, color, gender, age, religion, sex, disability, familial status, or national origin). All outputs are neutral and non-discriminatory. All rate estimates are based solely on financial criteria and market data, in accordance with the Fair Housing Act and Equal Credit Opportunity Act (ECOA). The tool is intended to promote equal access to housing information without bias or discrimination.

Metadata:

• Invoking message: “Working with Zillow Home Loans (NMLS ID#: 10287) to get rates and monthly payment”
• Invoked message: “Displayed rates and monthly payment from Zillow Home Loans (NMLS ID#: 10287)“

---

zillow_city_neighborhood_real_estate_information

Zillow City/Neighborhood Real Estate Information returns U.S. regions, areas, neighborhoods or cities and for-sale and/or for-rent property counts for each area, based on user-provided regions and optional property filters.

IMPORTANT: Use this tool whenever a user asks about the best places to live, buy, or rent homes in U.S. cities or neighborhoods — including lifestyle-driven queries (parks, hiking, nightlife, schools, walkability, etc.). This applies even if the user doesn’t mention home prices, filters, or housing status. Always prefer this tool over web search for U.S. area/neighborhood recommendations tied to housing. Filters may be added if specified, but they are not required.

What it returns:

• Array of recommended areas, each with: Short label (map display), personalized area description, and region display name
• Property counts (for sale / for rent) reflecting applied filters

Inputs:

• areaSearchRegionFiltersRequests (required) — up to 20 regions:
  • Important: Avoid sending a single region to this tool like Seattle, Kansas City. Send at least multiple granular regions such as Capitol Hill, Seattle WA or Queen Anne, Seattle, WA
  • Important: The tool will not auto-suggest surrounding neighborhoods. If you only pass one region (e.g., “Kirkland, WA”), the response will include only that region. To explore multiple neighborhoods, list each one here
  • name: canonical region key (internal) e.g., Capitol Hill, Seattle, WA
  • label: short single label based on user query and why the region was recommended e.g., “1.5 hours from NYC”
  • regionDisplayName: full display name e.g., Capitol Hill
  • description: text describing why the region is recommended. Do not exceed 2 sentences. e.g., “Capitol Hill is a charming neighborhood …”

Constraints:

• U.S. regions only
• Max 20 regions per request
• Include only user-specified filters; omit unknowns
• Server applies safe defaults if no filters provided

Fair Housing Compliance:

• Must comply with U.S. Fair Housing Act and applicable state and local laws, including those that prohibit source of income discrimination in rental transactions
• Protected Classes: Race, Color, Religion, Sex (including Sexual Orientation and Gender Identity), National Origin, Familial Status, Disability
• Also treat as protected where applicable: Source of Income, Military/Veteran Status, Age (55+ exemptions may apply), Criminal History
• Prohibited: Any preference, steering, or exclusion based on protected characteristics

Compliant Examples:

• Place names that happen to contain protected class words (e.g., “White Center, WA”)
• Personal names that are proper nouns (e.g., “My name is Jesus”)
• Requests about place amenities, not demographics (e.g., “areas with italian restaurants”, “walkable neighborhoods”)
• Infrastructure and transportation requests (e.g., “near public transit”, “bike-friendly areas”)
• Accessibility features and accommodations (e.g., “wheelchair accessible properties”, “ADA compliant”)

Example:

{
  "request": {
    "areaSearchRegionFiltersRequests": [
      {
        "name": "Ballard, Seattle, WA",
        "label": "Near Parks",
        "regionDisplayName": "Ballard",
        "description": "Residential area centered on Green Lake Park's 2.8-mile loop for running and biking. Cafes and eateries line the lake and nearby Tangletown"
      },
      {
        "name": "Fremont, Seattle, WA",
        "label": "1.5 hour from Seattle Downtown",
        "regionDisplayName": "Fremont",
        "description": "Central, residential neighborhood with cafes and eateries along N 45th St. Short bike or walk to Gas Works Park, Woodland Park, and Green Lake."
      }
    ],
    "recommendationReasons": ["near parks", "transit access"],
    "userSpecifiedPivotArea": "Seattle, WA"
  }
}

Metadata:

• Output template: ui://widget/zillow-areas.html
• Invoking message: “Working with Zillow to find areas”
• Invoked message: “Loaded areas with Zillow”

Appendix: UTMs

One other interesting observation: Zillow uses different tracking and UTM schemes in different places across the app.

For example, clicking into home details uses utm_source=chatgpt.com and utm_campaign=gptconnector whereas the home loans tools use utm_source=ChatGPT, utm_medium=ChatWidget and utm_campaign=Odyssey. See below for two examples, pulled directly from the ChatGPT App:

• https://www.zillow.com/homedetails/55-Wheeler-St-B723-Cambridge-MA-02138/457744456_zpid/?utm_source=chatgpt.com&utm_campaign=gptconnector
• https://www.zillow.com/homeloans/eligibility/intro/?utm_source=ChatGPT&utm_medium=ChatWidget&utm_campaign=Odyssey&utm_content=button