Northeast Deal Intel — MCP Server
Connect any MCP-compatible LLM (Claude, Cursor, Continue) to 14,000+ AI-scored commercial real estate deals and 100,000+ closed comps across the Northeast.
What You Can Do
Ask your LLM natural language questions like:
- "Find industrial deals in CT over 7% cap rate under $2M"
- "I have a $1.4M 1031 exchange closing in 45 days. What NNN retail fits?"
- "Compare the cap rate on this Hartford warehouse to recent comps"
- "Score this deal: 215 Main St, Windsor CT, $3.2M industrial, 7.8% cap"
- "Which CT deals have the highest sell probability right now?"
Tools Exposed
| Tool | Description | Tier Required |
|---|---|---|
search_deals | Find active listings by state, type, score, price, cap rate | Any |
get_deal | Full deal details + scoring breakdown + sell signal | Any |
search_comps | 100K+ closed transactions for benchmarking | agent_starter |
score_deal | Submit any deal for AI scoring | agent_pro |
get_market_benchmarks | Cap rate + PSF benchmarks by state/type | Any |
find_1031_candidates | Filtered search for exchange-ready deals | Any |
get_sell_signal | Sell probability for a specific listing | agent_starter |
get_market_summary | State-level market overview | Any |
Setup
1. Get an API Key
Sign up at northeastdealintel.com/agent-api.html
- Agent Starter ($49/mo) —
search_deals,get_deal,search_comps,get_sell_signal,get_market_summary - Agent Pro ($149/mo) — Everything above +
score_deal(submit any deal for AI scoring) - Agent Enterprise ($499/mo) — Full access + bulk exports + custom comp reports
2. Install dependencies
pip install mcp httpx
3. Configure Claude Desktop
Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or
%APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"northeast-deal-intel": {
"command": "python3",
"args": ["/path/to/ndi_mcp_server.py"],
"env": {
"NDI_API_KEY": "your_api_key_here"
}
}
}
}
Restart Claude Desktop. You'll see the NDI tools available in the tool picker.
4. Configure for other MCP clients
The server uses stdio transport (standard). Any MCP-compatible client works the same way — just point it at ndi_mcp_server.py with NDI_API_KEY in the environment.
Cursor: Add to .cursor/mcp.json in your project root.
Continue: Add to ~/.continue/config.json under mcpServers.
Example Conversations
1031 Exchange
"I sold a CT strip center for $1.8M and need a replacement property within 45 days. Find me NNN retail with cap rates above 7% priced between $1.5M and $2.2M."
Market Research
"What's the average cap rate for industrial in Hartford vs. Fairfield County right now? Show me the top 5 deals in each submarket."
Deal Underwriting
"Score this deal: 45 Industrial Dr, Wallingford CT. Asking $4.1M, 47,000 SF warehouse, single tenant NNN, 6.8% cap, 4 years remaining on lease."
Distress Hunting
"Find me deals in CT with high sell probability — I'm looking for motivated sellers. Focus on industrial and multifamily."
API Base URL
https://api.northeastdealintel.com
Override with env var: NDI_API_BASE=https://api.northeastdealintel.com