# PropData Address Autocomplete & AVM API > Real estate address typeahead with instant valuation, deal signals, market intelligence, and neighborhood demographics — built on 165M+ government-sourced US parcels. **Version:** 7.0.0 **Base URL:** `https://propdata-autocomplete.proptechusa.ai` **Provider:** PropTechUSA.ai — Saint Paul, MN **Support:** propdata@proptechusa.ai --- ## Table of Contents 1. [Overview](#overview) 2. [Authentication](#authentication) 3. [Rate Limits](#rate-limits) 4. [Endpoints](#endpoints) - [GET /autocomplete](#get-autocomplete) - [POST /select](#post-select) - [GET /health](#get-health) 5. [Response Reference](#response-reference) - [AutocompleteResult](#autocompleteresult) - [SpreadSignal](#spreadsignal) - [MarketIntel](#marketintel) - [NeighborhoodData](#neighborhooddata) 6. [AVM Engine](#avm-engine) 7. [Deal Signal Logic](#deal-signal-logic) 8. [Self-Improving Results](#self-improving-results) 9. [Error Codes](#error-codes) 10. [Code Examples](#code-examples) 11. [FAQ](#faq) --- ## Overview Most address autocomplete APIs return a street name and coordinates. This one returns the street name plus everything a real estate investor, developer, or acquisitions rep needs to make a decision — before the user finishes typing. **Every autocomplete result includes:** | Field | Description | |-------|-------------| | Address + parcel ID | Matched against 165M+ US parcels | | Property attributes | Beds, baths, sqft, year built | | Last sale | Price and date | | AVM | Automated valuation, 9-tier fallback chain | | AVM source | Exactly how the value was computed | | Confidence | `high / medium / medium-low / low` | | Deal signal | `undervalued / market / overimproved` vs ZIP median | | Market intel | ZHVI, HUD rent, gross yield, YoY appreciation, P/R ratio | | Demographics | Census ACS: income, population, vacancy, poverty, owner-occupied % | **Data sources:** - 165M+ county assessor parcel records (ArcGIS, government-sourced) - Zillow ZHVI (ZIP-level home values + appreciation) - HUD Fair Market Rents (2BR benchmark by county) - Census ACS 5-year estimates (neighborhood demographics) - US Census geocoder + OpenStreetMap (fallback for unmatched addresses) --- ## Authentication All requests require a RapidAPI key passed in the request header. ``` X-RapidAPI-Key: YOUR_RAPIDAPI_KEY X-RapidAPI-Host: propdata-autocomplete.proptechusa.ai ``` RapidAPI injects these headers automatically when you use the RapidAPI console or SDKs. If calling directly, include both headers on every request. --- ## Rate Limits | Plan | Requests/Month | Rate Limit | Overages | |------|---------------|------------|----------| | STARTER | 25,000 | — | None | | PRO | 100,000 | — | None | | ULTRA | 500,000 | — | None | **No per-call fees. Ever.** All plans are flat rate. If you hit your monthly quota, requests are blocked until reset or upgrade — you will never receive a surprise overage bill. **Caching:** Identical queries are cached at the edge for 60 minutes. Cached responses return in under 10ms and do not count against your quota. --- ## Endpoints --- ### GET /autocomplete Address typeahead search with AVM enrichment. **URL** ``` GET https://propdata-autocomplete.proptechusa.ai/autocomplete ``` **Query Parameters** | Parameter | Type | Required | Description | |-----------|------|----------|-------------| | `q` | string | ✅ | Address search query. Minimum 3 characters. | | `state` | string | ❌ | Filter to a US state. 2-letter code (e.g. `MN`, `TX`, `CA`). | | `limit` | integer | ❌ | Results to return. Min: 1, Max: 15, Default: 8. | **Request Example** ```http GET /autocomplete?q=123+Main+St&state=MN&limit=5 X-RapidAPI-Key: YOUR_KEY X-RapidAPI-Host: propdata-autocomplete.proptechusa.ai ``` **Response** Returns an array of [`AutocompleteResult`](#autocompleteresult) objects, ordered by similarity score descending. ```json [ { "parcel_id": "27053-ABC123", "address": "123 MAIN ST", "city": "Minneapolis", "state": "MN", "zip": "55401", "full_address": "123 MAIN ST, Minneapolis, MN, 55401", "score": 0.95, "source": "propdata", "bedrooms": 3, "bathrooms": 2, "sqft": 1450, "year_built": 1987, "last_sale_price": 285000, "last_sale_date": "2022-06-15", "avm": 312000, "avm_source": "assessed_x_ratio", "confidence": "medium", "improvement_value": 312000, "best_value": 312000, "zhvi_value": 308000, "zhvi_1yr_appreciation": 4.2, "zhvi_3yr_appreciation": 18.7, "spread_signal": { "ratio": 1.01, "label": "market", "deal_flag": false }, "market": { "zhvi_median": 308000, "rent_estimate": 1850, "gross_yield": 7.1, "price_to_rent": 167, "yoy_appreciation": 4.2, "unemployment": 3.8 }, "neighborhood": { "median_income": 68500, "per_capita_income": 38200, "population": 42300, "median_age": 34.2, "owner_occupied_pct": 52.1, "vacancy_rate": 5.3, "poverty_rate": 11.2, "college_degree_pct": 48.7, "median_rent": 1420 } } ] ``` **Response Headers** | Header | Description | |--------|-------------| | `X-Cache` | `HIT` if served from edge cache, `MISS` if freshly computed | | `X-API-Version` | Worker version (e.g. `v7.0.0`) | | `X-Powered-By` | `PropData 165M+` | --- ### POST /select Record a user address selection to improve future results. **URL** ``` POST https://propdata-autocomplete.proptechusa.ai/select ``` **When to call this:** Every time a user selects an address from your autocomplete dropdown. This is how the self-improving learned address system works — selected addresses are boosted in future results for all users in that market. Fire and forget. This call is async and should never block your UI. **Request Headers** ``` Content-Type: application/json X-RapidAPI-Key: YOUR_KEY ``` **Request Body** | Field | Type | Required | Description | |-------|------|----------|-------------| | `address` | string | ✅ | Street address of selected property | | `city` | string | ❌ | City name | | `state` | string | ❌ | 2-letter state code | | `zip` | string | ❌ | 5-digit ZIP code | | `parcel_id` | string | ❌ | Parcel ID from autocomplete result | | `avm` | integer | ❌ | AVM value from autocomplete result | | `avm_source` | string | ❌ | AVM source from autocomplete result | | `confidence` | string | ❌ | Confidence level from autocomplete result | | `source` | string | ❌ | Source field from autocomplete result | **Request Example** ```http POST /select Content-Type: application/json X-RapidAPI-Key: YOUR_KEY { "address": "123 MAIN ST", "city": "Minneapolis", "state": "MN", "zip": "55401", "parcel_id": "27053-ABC123", "avm": 312000, "avm_source": "assessed_x_ratio", "confidence": "medium", "source": "propdata" } ``` **Response** ```json { "ok": true, "learned": "123 main st minneapolis mn 55401" } ``` --- ### GET /health API status and feature flags. No authentication required. **URL** ``` GET https://propdata-autocomplete.proptechusa.ai/health ``` **Response** ```json { "status": "ok", "version": "v7.0.0", "parcels": "165M+", "features": { "learning": true, "zip_stats": false, "demographics": true, "deal_signals": true, "market_intel": true }, "ts": "2026-06-05T12:00:00.000Z" } ``` --- ## Response Reference --- ### AutocompleteResult | Field | Type | Nullable | Description | |-------|------|----------|-------------| | `parcel_id` | string | ✅ | Unique parcel identifier (`county_fips-parcel_id`) | | `address` | string | ✅ | Street address, uppercase | | `city` | string | ✅ | City name | | `state` | string | ✅ | 2-letter state code | | `zip` | string | ✅ | 5-digit ZIP code | | `full_address` | string | ❌ | Formatted full address string | | `score` | float | ❌ | Similarity score 0.0–1.0 (higher = stronger match) | | `source` | string | ❌ | Data source: `propdata`, `learned`, `census`, `nominatim` | | `bedrooms` | integer | ✅ | Number of bedrooms | | `bathrooms` | number | ✅ | Number of bathrooms | | `sqft` | integer | ✅ | Square footage of living area | | `year_built` | integer | ✅ | Year built | | `last_sale_price` | integer | ✅ | Most recent sale price in USD | | `last_sale_date` | string | ✅ | Most recent sale date (YYYY-MM-DD) | | `avm` | integer | ✅ | Automated valuation estimate in USD | | `avm_source` | string | ✅ | Which AVM tier produced this estimate | | `confidence` | string | ✅ | `high`, `medium`, `medium-low`, or `low` | | `improvement_value` | integer | ✅ | Alias for `avm` (backward compatibility) | | `best_value` | integer | ✅ | Alias for `avm` (backward compatibility) | | `zhvi_value` | integer | ✅ | Zillow Home Value Index for this ZIP | | `zhvi_1yr_appreciation` | float | ✅ | ZHVI 1-year appreciation % | | `zhvi_3yr_appreciation` | float | ✅ | ZHVI 3-year appreciation % | | `spread_signal` | object | ✅ | Deal signal vs ZIP median. See [SpreadSignal](#spreadsignal) | | `market` | object | ✅ | ZIP-level market intel. See [MarketIntel](#marketintel) | | `neighborhood` | object | ✅ | Census ACS demographics. See [NeighborhoodData](#neighborhooddata) | **Score ranges by source:** | Source | Typical Score | Description | |--------|--------------|-------------| | `propdata` | 0.70–1.00 | Native parcel database match | | `learned` | ~0.60 | Previously geocoded and learned | | `nominatim` | ~0.50 | OpenStreetMap geocoder fallback | | `census` | ~0.45 | US Census geocoder fallback | --- ### SpreadSignal Compares this property's AVM to its ZIP median (ZHVI). Null if ZIP median data is unavailable. | Field | Type | Description | |-------|------|-------------| | `ratio` | float | `property_avm / zip_zhvi_median` | | `label` | string | `undervalued`, `market`, or `overimproved` | | `deal_flag` | boolean | `true` if ratio < 0.7 | **Label thresholds:** | Ratio | Label | Meaning | |-------|-------|---------| | < 0.70 | `undervalued` | AVM is 30%+ below ZIP median — potential deal | | 0.70–1.30 | `market` | AVM is within normal range for this ZIP | | > 1.30 | `overimproved` | AVM is 30%+ above ZIP median — seller overspent | --- ### MarketIntel ZIP-level market intelligence. Null if ZIP is not in AVM cache. | Field | Type | Nullable | Description | |-------|------|----------|-------------| | `zhvi_median` | integer | ✅ | ZHVI median home value for this ZIP | | `rent_estimate` | integer | ✅ | HUD Fair Market Rent (2BR) in USD/month | | `gross_yield` | float | ✅ | Estimated gross rental yield % | | `price_to_rent` | float | ✅ | Price-to-rent ratio (lower = better for investors) | | `yoy_appreciation` | float | ✅ | Year-over-year home value appreciation % | | `unemployment` | float | ✅ | Local unemployment rate % | **Interpreting gross yield:** | Gross Yield | Signal | |-------------|--------| | > 8% | Strong cash flow market | | 5–8% | Moderate cash flow | | < 5% | Appreciation play, thin cash flow | --- ### NeighborhoodData Census ACS 5-year estimates. Null if ZIP has no Census ACS data. | Field | Type | Nullable | Description | |-------|------|----------|-------------| | `median_income` | integer | ✅ | Median household income in USD | | `per_capita_income` | integer | ✅ | Per capita income in USD | | `population` | integer | ✅ | Total population | | `median_age` | float | ✅ | Median age of residents | | `owner_occupied_pct` | float | ✅ | % of housing units that are owner-occupied | | `vacancy_rate` | float | ✅ | Housing vacancy rate % | | `poverty_rate` | float | ✅ | Poverty rate % | | `college_degree_pct` | float | ✅ | % of adults with a college degree | | `median_rent` | integer | ✅ | Median gross rent in USD/month | --- ## AVM Engine The AVM (Automated Valuation Model) uses a 9-tier fallback chain. Every result tells you exactly which tier produced the estimate and how confident to be. ``` Tier 1 → Recent sale (≤18 months) confidence: high Tier 2 → Assessor market value field confidence: high Tier 3 → Assessed value × state ratio confidence: medium Tier 4 → Improvement value × ratio confidence: medium Tier 5 → Older sale + 4%/yr appreciation confidence: medium-low Tier 6 → ZIP stats median confidence: medium-low Tier 7 → ZHVI/AVM ZIP cache (sqft-adjusted) confidence: low Tier 8 → State median confidence: low Tier 9 → null no data available ``` **avm_source values:** | Value | Tier | Description | |-------|------|-------------| | `recent_sale` | 1 | Sale within 18 months | | `market_value` | 2 | Assessor market value field | | `assessed_x_ratio` | 3 | Assessed value × state assessment ratio | | `improvement_x_ratio` | 4 | Improvement value × ratio (capped 1.4×) | | `sale_appreciated` | 5 | Older sale + 4%/yr compound appreciation | | `zip_median_recent` | 6 | ZIP recent sale median | | `zip_median_market` | 6 | ZIP market value median | | `zip_median_sale` | 6 | ZIP all-sale median | | `zhvi_*` | 7 | ZHVI/AVM cache baseline (sqft-adjusted if available) | | `state_median` | 8 | State-level median home value | | `learned` | — | Previously cached from user activity | **State assessment ratios** are sourced from IAAO and state DOE publications. For example: - South Carolina: 25× (4% primary residence assessment ratio) - Michigan: 2× (50% state equalized value) - Illinois: 10× (Cook County ~10% ratio) - California: 1× (Prop 13 — assessed ≠ market, use with caution) --- ## Deal Signal Logic The `spread_signal` compares the property's AVM to its ZIP-level ZHVI median. ``` ratio = property_avm / zip_zhvi_median ratio < 0.70 → undervalued → deal_flag: true ratio 0.70–1.30 → market → deal_flag: false ratio > 1.30 → overimproved → deal_flag: false ``` **Use cases:** - **Filter for deals:** `deal_flag === true` surfaces properties priced 30%+ below their ZIP median — a strong signal for wholesalers and fix-and-flip investors - **Seller counseling:** `overimproved` means the owner has spent more improving the property than the market rewards — useful for listing agents setting price expectations - **Portfolio screening:** Screen large address lists by `label` to bucket into deal / hold / pass **Important:** The spread signal is a relative indicator, not an appraisal. A `deal_flag: true` result means the AVM is below the ZIP median — it does not guarantee the property is actually available, distressed, or priced below its true market value. Always verify with additional due diligence. --- ## Self-Improving Results The API learns from usage. Here is how it works: **Address learning:** When the autocomplete falls back to the Census geocoder or Nominatim (because the address wasn't in the PropData database), that result is automatically saved to the learned address cache. The next time anyone searches for that address, it returns from the learned cache — faster, with any previously computed AVM attached. **Selection boosting:** When you call `POST /select` after a user picks an address, that address gets its `select_count` incremented. Learned addresses are sorted by `select_count` descending — the most-selected addresses in your market surface first. **Best practice:** ```javascript // After user selects from dropdown async function onAddressSelect(result) { // Don't await — fire and forget fetch('https://propdata-autocomplete.proptechusa.ai/select', { method: 'POST', headers: { 'Content-Type': 'application/json', 'X-RapidAPI-Key': YOUR_KEY }, body: JSON.stringify({ address: result.address, city: result.city, state: result.state, zip: result.zip, parcel_id: result.parcel_id, avm: result.avm, avm_source: result.avm_source, confidence: result.confidence, source: result.source }) }); } ``` --- ## Error Codes | Status | Code | Description | |--------|------|-------------| | 200 | — | Success | | 400 | Bad Request | Missing or invalid parameters. Check `error` field for details. | | 404 | Not Found | Endpoint does not exist | | 405 | Method Not Allowed | Wrong HTTP method for this endpoint | | 429 | Too Many Requests | Rate limit exceeded. Check `retry_after` field. | **Error response shape:** ```json { "error": "q must be at least 3 characters", "status": 400 } ``` **Rate limit response shape:** ```json { "error": "Rate limit exceeded", "limit": 10, "retry_after": "60s" } ``` --- ## Code Examples ### JavaScript / Fetch ```javascript const searchAddresses = async (query, state = null) => { const params = new URLSearchParams({ q: query, limit: '8' }); if (state) params.set('state', state); const response = await fetch( `https://propdata-autocomplete.proptechusa.ai/autocomplete?${params}`, { headers: { 'X-RapidAPI-Key': 'YOUR_RAPIDAPI_KEY', 'X-RapidAPI-Host': 'propdata-autocomplete.proptechusa.ai' } } ); if (!response.ok) { const err = await response.json(); throw new Error(err.error); } return response.json(); // AutocompleteResult[] }; // Usage const results = await searchAddresses('123 Main St', 'MN'); results.forEach(r => { console.log(`${r.full_address}`); console.log(` AVM: $${r.avm?.toLocaleString()} (${r.avm_source}, ${r.confidence})`); console.log(` Deal: ${r.spread_signal?.label} (ratio: ${r.spread_signal?.ratio})`); console.log(` Yield: ${r.market?.gross_yield}%`); }); ``` --- ### React Autocomplete Component ```jsx import { useState, useCallback } from 'react'; const RAPIDAPI_KEY = 'YOUR_RAPIDAPI_KEY'; const BASE_URL = 'https://propdata-autocomplete.proptechusa.ai'; function AddressAutocomplete({ onSelect, stateFilter = null }) { const [query, setQuery] = useState(''); const [results, setResults] = useState([]); const [loading, setLoading] = useState(false); const search = useCallback(async (q) => { if (q.length < 3) { setResults([]); return; } setLoading(true); try { const params = new URLSearchParams({ q, limit: '8' }); if (stateFilter) params.set('state', stateFilter); const res = await fetch(`${BASE_URL}/autocomplete?${params}`, { headers: { 'X-RapidAPI-Key': RAPIDAPI_KEY, 'X-RapidAPI-Host': 'propdata-autocomplete.proptechusa.ai' } }); const data = await res.json(); setResults(Array.isArray(data) ? data : []); } catch (_) { setResults([]); } finally { setLoading(false); } }, [stateFilter]); const handleSelect = async (result) => { setQuery(result.full_address); setResults([]); onSelect(result); // Fire-and-forget selection learning fetch(`${BASE_URL}/select`, { method: 'POST', headers: { 'Content-Type': 'application/json', 'X-RapidAPI-Key': RAPIDAPI_KEY, 'X-RapidAPI-Host': 'propdata-autocomplete.proptechusa.ai' }, body: JSON.stringify({ address: result.address, city: result.city, state: result.state, zip: result.zip, parcel_id: result.parcel_id, avm: result.avm, avm_source: result.avm_source, confidence: result.confidence, source: result.source }) }); }; return (

<input value={query} onChange={e => { setQuery(e.target.value); search(e.target.value); }} placeholder="Search address..." style={{ width: '100%', padding: '8px 12px', fontSize: 14 }} /> {loading &&

Searching...

} {results.length > 0 && (

handleSelect(r)} style={{ padding: '10px 12px', cursor: 'pointer', borderBottom: '1px solid #f0f0f0' }} >
{r.full_address}

{r.avm ? `AVM: $${r.avm.toLocaleString()} · ` : ''} {r.spread_signal?.label && `${r.spread_signal.label} · `} {r.market?.gross_yield ? `${r.market.gross_yield}% yield` : ''}

)}

); } export default AddressAutocomplete; ``` --- ### Python ```python import requests RAPIDAPI_KEY = "YOUR_RAPIDAPI_KEY" BASE_URL = "https://propdata-autocomplete.proptechusa.ai" HEADERS = { "X-RapidAPI-Key": RAPIDAPI_KEY, "X-RapidAPI-Host": "propdata-autocomplete.proptechusa.ai" } def search_addresses(query: str, state: str = None, limit: int = 8) -> list: params = {"q": query, "limit": limit} if state: params["state"] = state resp = requests.get(f"{BASE_URL}/autocomplete", params=params, headers=HEADERS) resp.raise_for_status() return resp.json() def record_selection(result: dict) -> None: payload = { "address": result.get("address"), "city": result.get("city"), "state": result.get("state"), "zip": result.get("zip"), "parcel_id": result.get("parcel_id"), "avm": result.get("avm"), "avm_source": result.get("avm_source"), "confidence": result.get("confidence"), "source": result.get("source"), } requests.post(f"{BASE_URL}/select", json=payload, headers=HEADERS) # Usage results = search_addresses("123 Main St", state="MN") for r in results: print(f"{r['full_address']}") print(f" AVM: ${r['avm']:,} ({r['avm_source']}, {r['confidence']})") if r.get("spread_signal"): sig = r["spread_signal"] print(f" Deal signal: {sig['label']} (ratio: {sig['ratio']})") if r.get("market"): mkt = r["market"] print(f" Gross yield: {mkt['gross_yield']}%") print(f" YoY appreciation: {mkt['yoy_appreciation']}%") if r.get("neighborhood"): nbr = r["neighborhood"] print(f" Median income: ${nbr['median_income']:,}") print(f" Vacancy rate: {nbr['vacancy_rate']}%") print() ``` --- ### Node.js / Axios ```javascript const axios = require('axios'); const client = axios.create({ baseURL: 'https://propdata-autocomplete.proptechusa.ai', headers: { 'X-RapidAPI-Key': 'YOUR_RAPIDAPI_KEY', 'X-RapidAPI-Host': 'propdata-autocomplete.proptechusa.ai' } }); async function searchAddresses(query, state = null, limit = 8) { const params = { q: query, limit }; if (state) params.state = state; const { data } = await client.get('/autocomplete', { params }); return data; } async function recordSelection(result) { await client.post('/select', { address: result.address, city: result.city, state: result.state, zip: result.zip, parcel_id: result.parcel_id, avm: result.avm, avm_source: result.avm_source, confidence: result.confidence, source: result.source }); } // Screen a list of addresses for deals async function findDeals(addresses) { const deals = []; for (const addr of addresses) { const results = await searchAddresses(addr); const deal = results.find(r => r.spread_signal?.deal_flag === true); if (deal) deals.push(deal); } return deals; } ``` --- ## FAQ **Q: How fresh is the data?** County assessor records are updated nightly from ArcGIS FeatureServer endpoints. ZHVI data is updated monthly. Census ACS data is updated annually. Sale records reflect what counties have published — lag varies by county from days to weeks. **Q: What happens when there is no AVM?** `avm` returns `null` and `avm_source` returns `null`. This is rare — the 9-tier fallback chain reaches state median before returning null, and every US state has a median value. Null only occurs for non-residential parcels or addresses with no state association. **Q: Why does the same address sometimes return different AVM values?** The AVM tier used depends on what data is available for that parcel. A parcel that had a recent sale will use `recent_sale` (high confidence). The same parcel six months later may have no data in the recent-sale window and fall back to `assessed_x_ratio`. The `avm_source` field always tells you exactly which tier was used. **Q: What is the `score` field?** Similarity score between the search query and the matched address. PropData native results score 0.7–1.0. Learned addresses score approximately 0.6. Geocoder fallbacks score 0.45–0.5. Use this to rank results or filter out weak matches. **Q: Can I filter by city instead of state?** Not directly — use the `q` parameter to include the city name in your search string (e.g. `q=123+Main+St+Chicago`) and the matching algorithm will prefer results in that city. Use `state=IL` to further constrain. **Q: What does `source: learned` mean?** The address was not in the PropData native database but was previously found via a geocoder fallback and saved to the learned cache. Learned results may have less property attribute data but benefit from any AVM that was computed when they were first learned. **Q: Is this compliant with Fair Housing law?** The API returns demographic data from Census ACS. It is your responsibility to use this data in compliance with the Fair Housing Act, Equal Credit Opportunity Act, and all applicable federal, state, and local anti-discrimination laws. Do not use neighborhood demographics to make lending, rental, or purchase decisions based on protected characteristics. See our Terms of Service for full details. **Q: Do you have a full property records API?** Yes. PropData — Real Estate Market Intelligence API includes full property lookup, market intelligence, rent estimates, comps, preforeclosure leads, skip trace, state-level signals, and more. 166M+ parcels, 16 live data sources, $49/mo flat. → [View on RapidAPI](https://rapidapi.com/propdata-default/api/propdata-real-estate-market-intelligence) --- *PropData Address Autocomplete & AVM API — v7.0.0* *Built by PropTechUSA.ai — Saint Paul, MN* *propdata@proptechusa.ai | https://propdata.proptechusa.ai*