Use PropSWOT's Maintenance Intelligence Inside Claude, Gemini, and Your AI Tools

Claude Desktop

Live — available now

Claude Desktop (Anthropic) natively supports MCP servers via a local config file. This is the recommended path for property managers who already use Claude Desktop on their computer.

macOS

Config file: ~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "propswot": {
      "url": "https://pm.propswot.com/api/mcp",
      "headers": {
        "X-MCP-Key": "ps_live_YOUR_KEY_HERE"
      }
    }
  }
}

Windows

Config file: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "propswot": {
      "url": "https://pm.propswot.com/api/mcp",
      "headers": {
        "X-MCP-Key": "ps_live_YOUR_KEY_HERE"
      }
    }
  }
}

Restart Claude Desktop. You'll see a "propswot" entry in the tools panel. Try these prompts to verify:

• "What maintenance issues need my attention today?"
• "Show me open work orders at 123 Main St."
• "What HVAC vendors do I have on file?"
• "Generate a self-help link in Spanish for work order #42."

Claude.ai Projects

Live — no installation required

Claude.ai Projects (Anthropic's web product) supports remote MCP servers directly from the browser — no config file or desktop app installation needed. This is the lowest-friction path for property managers who work in the browser.

1. Open claude.ai and navigate to Projects (left sidebar).
2. Open or create a Project for property management work.
3. Click the Settings icon (gear) inside the Project.
4. Under Tools, click Add MCP server.
5. Enter the server URL: https://pm.propswot.com/api/mcp
6. Add a custom header — name: X-MCP-Key, value: your API key.
7. Save. PropSWOT tools will now be available to Claude in every conversation inside that Project.

ChatGPT (GPT Actions)

Live — available now

ChatGPT Plus and Enterprise users can add PropSWOT as a custom GPT with Actions. The PM creates a private "PropSWOT Assistant" GPT once, and from then on they just type natural language questions in ChatGPT — no dashboard needed. Enterprise admins can publish the GPT org-wide so every PM on the team gets it pre-configured.

Setup (5 minutes)

1. Generate a PropSWOT API key at Dashboard → MCP / API Keys → New key. Name it "ChatGPT", choose Read or Write scope, and copy the key shown once.
2. In ChatGPT, go to My GPTs → Create. Name it "PropSWOT Assistant". Add a system prompt like: "You help me manage maintenance work orders for my properties. Always use PropSWOT tools to look up real data before answering."
3. Click Add actions. Choose Import from URL and paste:

   https://pm.propswot.com/openapi.yaml

   ChatGPT will load all 16 PropSWOT tools automatically.
4. Under Authentication, select API Key, Auth Type: Bearer, and paste your PropSWOT API key.
5. Click Save. Your "PropSWOT Assistant" GPT is ready.

Try these prompts

The OpenAPI spec is always current at https://pm.propswot.com/openapi.yaml. Use Import from URL to pick up new tools automatically when PropSWOT adds them.

Google Gemini

Via Gemini API — adapter required

Google Gemini does not natively speak the Model Context Protocol as of 2026. However, Gemini 2.0's function calling capability and MCP are structurally compatible — both use JSON-described tool schemas and structured call/response cycles. A thin adapter bridges the two.

The adapter pattern: declare PropSWOT's MCP tools as Gemini FunctionDeclaration objects, pass them in the Gemini API request, intercept function_call responses, forward them to the PropSWOT MCP server, and return the results as function_response parts.

import { GoogleGenerativeAI } from "@google/generative-ai";
import { Client } from "@modelcontextprotocol/sdk/client/index.js";
import { StreamableHTTPClientTransport } from "@modelcontextprotocol/sdk/client/streamableHttp.js";

// 1. Connect to PropSWOT MCP
const mcp = new Client({ name: "gemini-adapter", version: "1.0.0" });
await mcp.connect(new StreamableHTTPClientTransport(
  new URL("https://pm.propswot.com/api/mcp"),
  { headers: { "X-MCP-Key": process.env.PROPSWOT_MCP_KEY } }
));

// 2. Fetch tool list and convert to Gemini FunctionDeclarations
const { tools } = await mcp.listTools();
const geminiTools = tools.map(t => ({
  name: t.name,
  description: t.description,
  parameters: t.inputSchema,
}));

// 3. Call Gemini with PropSWOT tools available
const genai = new GoogleGenerativeAI(process.env.GEMINI_API_KEY);
const model = genai.getGenerativeModel({
  model: "gemini-2.0-flash",
  tools: [{ functionDeclarations: geminiTools }]
});
const chat = model.startChat();
let response = await chat.sendMessage("What work orders need attention today?");

// 4. Handle function_call → MCP → function_response loop
while (response.functionCalls()?.length) {
  const call = response.functionCalls()[0];
  const mcpResult = await mcp.callTool(call.name, call.args);
  response = await chat.sendMessage([{
    functionResponse: { name: call.name, response: mcpResult }
  }]);
}

console.log(response.text());

PropSWOT publishes its full tool schema at https://pm.propswot.com/.well-known/mcp.json — use this to auto-generate the Gemini function declarations rather than writing them by hand.