Vexcel MCP

Connect Vexcel to Your AI Tool

Overview

MCP (Model Context Protocol) is an open standard that lets AI tools connect directly to external data sources and services. Rather than copying data into a chat window, the AI queries live data on your behalf and incorporates the results directly into its responses. The AI can then reason over those results, combining them with other data, making follow-up requests, or shaping its response to your specific question.

Vexcel’s MCP draws from the same data available through the Vexcel API, covering much of its imagery, elevation, property, damage, and embeddings capabilities, letting you access that data by simply asking your AI in plain language. This guide walks you through connecting the Vexcel MCP server to your AI tool of choice.

What Vexcel Offers

Vexcel provides tools across several main product categories:

• Orthomosaic Imagery: High-resolution top-down aerial photos, including disaster response (graysky) collections
• Oriented Imagery: Oblique and nadir aerial photos
• Digital Elevation Models: DSM and DTM data representing building surfaces and bare terrain, in meters
• Property & Structure Metadata: Roof conditions, building footprints, pools, solar panels, and more
• Damage Assessment: Post-disaster property-level analysis with structure damage classifications
• Vexcel Embeddings: Similarity search across imagery using text, image, or location-based queries
• Admin Tools: Query API health status, check usage, and view your account profile

Step-by-Step Setup

You’ll need two things before you start:

  • Your Vexcel API key provided by Vexcel support
  • The Vexcel MCP server URL: https://api.vexcelgroup.com/v2/mcp

Choose the path that matches how you’re using your AI tool:

Option A: AI Chat App

Most AI chat applications (such as Claude, ChatGPT, or Gemini) support MCP servers through a settings, connectors, or integrations panel. The exact steps vary by app, but the general flow is:

  1. Open your app’s Settings
  2. Find a section labeled MCP Servers, Connectors, or Integrations
  3. Add a new server and enter the Vexcel MCP URL
  4. Authenticate using one of the following methods. Your app will determine which applies:
    • Prompted: Enter your Vexcel API key when the app asks for authentication
    • URL parameter: Append your key directly to the server URL: https://api.vexcelgroup.com/v2/mcp?api_key=YOUR_API_KEY
  5. Save and start a new conversation
  6. Consult your app’s documentation if you can’t find the MCP settings.

Option B: Coding Environment or CLI

Most developer tools (such as Claude Code, Cursor, or VS Code extensions) register MCP servers via a JSON config file. Add the following entry to your tool’s MCP configuration:

{  
  "mcpServers": {
    "vexcel": {
      "url": "https://api.vexcelgroup.com/v2/mcp",
      "type": "http",
      "headers": {
        "X-API-KEY": "YOUR_API_KEY"
      }
    }
  }

Replace YOUR_API_KEY with your Vexcel API key. The config file location varies by tool. Check your tool’s documentation for where to find it. Save the file and restart your tool

Verify the Connection
Test with a simple query:

“Search for properties near Boulder, Colorado using Vexcel”

If you get back Vexcel data, you’re all set. If not, see Troubleshooting below.

Using Vexcel with Your AI Tool

Once connected, you can ask natural-language questions. These examples are just a starting point. The best way to explore what’s possible is to start asking.

Property Queries

“Find properties with pools within 500 meters of POINT(-105.270 40.015)”

Imagery Requests

“Get the latest aerial photo for this location: POINT(-122.419 37.775)”

Damage Analysis

“Show me hurricane-damaged properties in Miami with severe roof damage”

Natural Language Search

“Find properties that visually resemble a commercial greenhouse facility”

Complex Multi-Step Query

“Pull structure metadata for properties within 1 km of POINT(-104.82 38.83) near Colorado Springs. Identify properties with less than 50% defensible space clearance. For those, also retrieve tree cover percentage and roof material type — flag any with combustible materials such as wood shake. Rank results by overall fire risk.”

Helpful Tips

  • Use WKT format for coordinates: POINT(longitude latitude) or POLYGON((...coordinates...))
  • Be specific about layers if you know them: “urban” for cities, “wide-area” for regional coverage, “graysky” for disaster response
  • Ask your AI to explain any Vexcel terminology you’re unfamiliar with
  • Specify a date range when you need recent imagery. Collections update over time, so asking for data “from the last 12 months” or “after [date]” narrows results to what’s most relevant
  • Combine data types in a single query. You can request imagery and property metadata together, e.g. “Show me aerial photos and roof condition data for this address”
  • Check coverage before diving deep. If you’re unsure whether Vexcel has data for an area, ask your AI to query coverage first before requesting detailed extracts or metadata

Troubleshooting

Your AI tool doesn’t recognize Vexcel commands:

  • Double-check the MCP server URL is correct
  • Verify your Vexcel account is active
  • Try removing and re-adding the server
  • Restart your app or refresh your browser

MCP connects but Vexcel tools aren’t showing up:

  • Some AI apps require you to explicitly enable or activate an MCP server after adding it — check that the Vexcel server is toggled on

Authentication errors:

  • Confirm you entered the correct API key
  • Make sure there are no extra spaces or characters around your API key when pasting it in
  • Check if your API key has expired (contact Vexcel)

No data returned:

  • Verify Vexcel has coverage for your area of interest
  • Check your WKT geometry is valid and uses the right coordinate system (usually EPSG:4326)
  • Confirm your API key has access to the product you’re querying. Imagery, damage, embeddings, and other products may require separate entitlements
  • Try a broader search with fewer filters

Slow or incomplete responses:

  • Large area requests can return a lot of data. Try narrowing your geometry or adding filters to reduce result size

Still stuck? Contact Vexcel support at support@vexcelgroup.com