How to Integrate Retool with CoStar

CoStar does not provide self-serve API access. There is no developer portal, no free tier, no trial access, and no API keys available to standard CoStar subscribers. The CoStar API is a separate enterprise product that requires a dedicated contract. The application process works as follows. Contact CoStar's sales team through costar.com or your existing CoStar account manager. Describe your use case: what data you need (property search, market analytics, comparables, tenant data, sale transactions), the approximate query volume, and how you plan to use the data (internal analytics tool, CRM enrichment, portfolio management platform). CoStar evaluates applications based on use case legitimacy and existing relationship — teams with an existing CoStar subscription have a better chance of approval. Upon approval, CoStar provides API documentation (typically a PDF or private portal), API credentials (API key, username/password, or OAuth client credentials depending on the API version), and a sandbox environment for testing. The production API is typically available at https://api.costar.com or a similar enterprise endpoint. CoStar's API versions vary — ask explicitly about the current API version and request the OpenAPI specification if available. If you do not yet have CoStar API access and need commercial real estate data for a Retool prototype or internal tool, consider these alternatives while pursuing approval: - ATTOM Data Solutions: property data API with self-serve access (attomdata.com) - Regrid: parcel data and property boundaries API - Estated: property data API with address-based lookups - CoStar's own export features: manually export comparables to CSV and import to Retool Database or a PostgreSQL table as a stopgap For teams that definitely need CoStar's specific data and scale, pursuing the enterprise API agreement is the only path to a live integration.

With your CoStar API credentials, configure a REST API Resource in Retool. Navigate to the Resources tab in the left sidebar and click Add Resource. Select REST API from the resource type list. Name the resource 'CoStar API' or 'CoStar Production'. In the Base URL field, enter the CoStar API base URL provided in your documentation (e.g., https://api.costar.com or https://api.costar.com/v1 — use the versioned base URL if CoStar specifies one). For authentication: CoStar's API authentication method varies by contract and API version. Common patterns include: - API Key: if CoStar provided an API key, add it as a header in the resource settings. Go to Headers → Add Header. Set header name to X-Api-Key or Authorization (check your documentation), and set the value to your API key. Mark it as a secret by referencing a Retool configuration variable: {{ configVars.COSTAR_API_KEY }}. - HTTP Basic Authentication: if CoStar provided a username and password, select Basic Auth in the Authentication dropdown. Enter the username and password. Retool will base64-encode them and send as an Authorization header. - OAuth 2.0 Client Credentials: if CoStar uses OAuth, select OAuth 2.0 and enter the client ID, client secret, and token URL provided by CoStar. In Settings → Configuration Variables, create a variable named COSTAR_API_KEY (or COSTAR_USERNAME and COSTAR_PASSWORD for basic auth). Mark it as a secret so it's never exposed to Retool's frontend. Reference it in your resource's authentication configuration. Click Save Changes. Unlike database resources, Retool cannot automatically test REST API connectivity without a specific endpoint to call — the save will succeed regardless of whether the credentials are correct. You'll verify the connection by running a test query in the next step.

Open a Retool app and create a new query. Select your CoStar REST API resource. The query editor will show fields for Method (GET/POST/PUT/DELETE), URL Path (relative to the base URL), URL Parameters, Headers, and Body. Start with a basic property search query. CoStar's property search endpoint typically accepts geographic and property type filters as query parameters or a JSON body (check your specific API documentation). Create a query named searchProperties. Set Method to GET (or POST if the endpoint requires a request body for search filters). Set the URL Path to the property search endpoint from your documentation (e.g., /property/search or /v1/properties). Add the required query parameters referencing Retool component values. Run the query. The response panel below the query editor shows the raw JSON response from CoStar. Inspect the response structure carefully — CoStar responses typically wrap results in a metadata envelope with pagination information, total count, and a results array. Create a JavaScript transformer to extract and reshape the results array into a flat format suitable for Retool's Table component. CoStar responses often include nested objects (address details, building specifications, market identifiers) that need to be flattened for tabular display. For market analytics endpoints: CoStar typically provides time-series data as arrays of period objects. These are well-suited for Retool's Chart component — transform the series data into the x/y format that Retool's Chart expects.

1// CoStar property search query configuration
2// Method: GET (or POST depending on CoStar API version)
3// URL Path: /property/search
4// URL Parameters:
5{
6  "marketId": "{{ marketSelect.value }}",
7  "propertyType": "{{ propertyTypeSelect.value }}",
8  "minSize": "{{ minSizeInput.value || '' }}",
9  "maxSize": "{{ maxSizeInput.value || '' }}",
10  "pageNumber": "{{ Math.floor(table1.pageIndex) + 1 }}",
11  "pageSize": "{{ table1.pageSize }}"
12}
13
14// Transformer to reshape CoStar property search results:
15const response = data;
16const properties = response.properties || response.results || response.data || [];
17
18return properties.map(prop => ({
19  propertyId: prop.propertyId || prop.id,
20  address: prop.address?.deliveryLine || prop.streetAddress,
21  city: prop.address?.city || prop.city,
22  market: prop.market?.name || prop.marketName,
23  submarket: prop.submarket?.name || prop.submarketName,
24  propertyType: prop.propertyType,
25  buildingSizeSF: prop.buildingSize?.toLocaleString(),
26  yearBuilt: prop.yearBuilt,
27  vacancyRate: prop.vacancy ? `${(prop.vacancy * 100).toFixed(1)}%` : 'N/A',
28  askingRent: prop.askingRentPerSF ? `$${prop.askingRentPerSF.toFixed(2)}/SF` : 'N/A',
29  lastSaleDate: prop.lastSaleDate ? new Date(prop.lastSaleDate).toLocaleDateString() : 'N/A'
30}));

With working CoStar queries, assemble the dashboard layout in Retool's visual editor. The typical CRE analytics dashboard structure for a Retool + CoStar integration includes: filter controls at the top (market selector, property type, size range, date range), KPI cards showing current market conditions, a Chart component with trend data, and a Table with detailed property or comparable records. For the filter controls: add a Select component for market selection, populate its options from a getMarkets query. Add a second Select for property type with hardcoded options (Office, Industrial, Retail, Multifamily, Mixed-Use). Add a DateRangePicker for transaction or listing date ranges. All filter components should trigger your CoStar queries automatically when changed — configure this in each query's Settings: enable 'Run query automatically when inputs change'. For KPI cards: Retool's Statistic component (or a styled Container with Text components) displays single metric values. Create a transformer that extracts the key metrics from your market statistics query response: current vacancy rate, net absorption, asking rent, construction pipeline. Bind each metric to a Statistic component. For the trend chart: CoStar market statistics typically return quarterly or monthly time-series data. Transform this into an array of { x: 'Q1 2024', vacancy: 8.5, rent: 42.50 } objects and bind to a Chart component configured as a multi-series line chart. For combining CoStar data with internal portfolio data: create a second query targeting your internal PostgreSQL or other database resource. Use a JavaScript transformer query to join the CoStar results with the internal database results on a shared property identifier (address, property ID, or a custom field your team maintains). For complex integrations with multiple data sources and sophisticated merge logic, RapidDev's team can help architect your complete Retool solution.

1// Market statistics transformer — format CoStar time-series data for Chart component
2// Input: data (CoStar market statistics response)
3const stats = data?.statistics || data?.timeSeries || data?.periods || [];
4
5return stats.map(period => ({
6  quarter: period.periodLabel || period.quarter || period.period,
7  vacancyRate: period.vacancyRate ? parseFloat((period.vacancyRate * 100).toFixed(2)) : null,
8  askingRent: period.askingRentPerSF ? parseFloat(period.askingRentPerSF.toFixed(2)) : null,
9  netAbsorption: period.netAbsorption || null,
10  newSupply: period.completions || period.newSupply || null
11})).sort((a, b) => {
12  // Sort chronologically by quarter string
13  return a.quarter < b.quarter ? -1 : a.quarter > b.quarter ? 1 : 0;
14});

CoStar's API enforces rate limits and returns paginated results for large data sets. Handling these correctly ensures your Retool dashboard is robust and doesn't fail under normal production usage. For pagination: CoStar typically returns a page of results with metadata indicating the total count and whether more pages exist. Check the response structure for fields like totalCount, pageCount, hasNextPage, or nextPageToken. In Retool, bind Table component pagination to your query's page and pageSize parameters. When the user advances to the next table page, Retool re-runs the query with the updated page parameter. For rate limits: CoStar's enterprise API rate limits are defined in your API agreement — these are typically measured in requests per minute or requests per day. Implement query caching in Retool for data that doesn't change frequently (market statistics are usually updated weekly or monthly). Enable query caching in the query's Advanced settings and set the cache duration appropriately (e.g., 3600 seconds = 1 hour for market data, 300 seconds for property listings). For error handling: CoStar API errors return standard HTTP error codes. Build error state handling in your Retool app: in each query's Event Handlers, add an On Failure handler that shows a notification with the error message. Common CoStar errors include 401 Unauthorized (expired or invalid credentials), 403 Forbidden (the endpoint is not included in your API access contract), 429 Too Many Requests (rate limit exceeded), and 400 Bad Request (invalid parameter values or missing required fields). For monitoring API usage: consider building a simple usage tracking Retool app that logs API calls to a Retool Database table (timestamp, endpoint, response time, status code) via a JavaScript query that runs after each CoStar API query. This helps identify which Retool apps are consuming the most CoStar API quota.

1// Pagination configuration for CoStar queries
2// URL Parameters (add to property search and comparables queries):
3{
4  "pageNumber": "{{ table1.pageIndex + 1 }}",
5  "pageSize": "{{ table1.pageSize }}",
6  "marketId": "{{ marketSelect.value }}",
7  "propertyType": "{{ propertyTypeSelect.value }}"
8}
9
10// Transformer to extract pagination metadata from CoStar response:
11const response = data;
12// Return structured object with items and pagination info
13// Bind table1.data to resultTransformer.data.items
14// Bind table1.totalCount to resultTransformer.data.totalCount
15return {
16  items: response.properties || response.results || [],
17  totalCount: response.totalCount || response.total || 0,
18  currentPage: response.pageNumber || 1,
19  pageCount: response.pageCount || Math.ceil((response.totalCount || 0) / (response.pageSize || 25))
20};