nj-realtor-mcp
A production-grade MCP server enabling Claude to perform comprehensive NJ real estate workflows including property search, valuation, neighborhood intelligence, investment analysis, and agent tools via 20 tools and 15+ data sources.
README
NJ Realtor MCP — Frontier Edition
A production-grade Model Context Protocol server for NJ real estate agents.
20 tools. 15+ data sources. Crawl4AI web crawling. Deep research mode. Full NJ-specific intelligence.
What This Does
Claude gains 20 native tools covering the full NJ real estate workflow:
| Category | Tools |
|---|---|
| Search & Discovery | search_listings, similar_homes, foreclosure_radar |
| Valuation | generate_cma, price_history, tax_assessment_lookup |
| Intelligence | neighborhood_report, school_district_report, flood_risk_report, commute_analysis |
| Investment | investment_analysis, rental_comps, mortgage_calculator |
| Agent Tools | draft_listing, schedule_showing, client_match, send_client_alert |
| Market Data | market_pulse, permit_lien_lookup |
| Research | deep_research (orchestrated multi-step AI research) |
Data Sources (Full Map)
MLS / Listing Data
| Source | What | How |
|---|---|---|
| GSMLS (Garden State MLS) | NJ's largest MLS — Bergen, Essex, Morris, Union, Passaic, Hudson, Somerset, Sussex, Warren | RESO Web API (OAuth2) |
| Bright MLS | South Jersey — Camden, Burlington, Atlantic, Cape May, Cumberland, Salem, Gloucester | Bright MLS API |
| Zillow | Zestimate, price history, public listing details | Crawl4AI (Playwright) |
| Redfin | Market trends, sold data CSVs, DOM stats | CSV export + Crawl4AI |
| Realtor.com | Competing listings, agent data | Crawl4AI |
Public Records (NJ-Specific)
| Source | What | How |
|---|---|---|
| NJ MOD-IV | Statewide property tax assessment database (all 565 municipalities) | Annual bulk download + API |
| NJ Courts Portal | Lis pendens, tax liens, judgments, foreclosures | Crawl4AI (PACER-style search) |
| NJ DEP GIS | Flood hazard delineations, wetlands, brownfields, contamination sites | ArcGIS REST API |
| NJ Flood Mapper | FEMA flood zones, base flood elevation, advisory maps | Crawl4AI |
| County Clerk portals | Deed records, mortgage filings (20 NJ counties each have own portal) | Crawl4AI per county |
| NJ OpenData | Municipal boundaries, zoning, permits, demographics | Socrata API (data.nj.gov) |
Neighborhood Intelligence
| Source | What | How |
|---|---|---|
| US Census ACS | Income, age, education, race, housing tenure (block group level) | Census API (free, key required) |
| GreatSchools | School ratings 1-10, test scores, enrollment | GreatSchools API |
| NJ DOE Report Card | District Factor Group (DFG), NJSLA scores, graduation rates | NJ OpenData / Crawl4AI |
| Walk Score | Walkability, transit score, bike score | Walk Score API |
| SpotCrime | Neighborhood crime incidents (0.5mi radius) | Crawl4AI / SpotCrime API |
| Google Places | Nearby grocery, restaurants, hospitals, parks, gyms | Google Maps Platform |
Market & Financial
| Source | What | How |
|---|---|---|
| FRED (Federal Reserve) | 30yr/15yr mortgage rates, housing price index, CPI | FRED API (free) |
| ATTOM Data | Property deed/mortgage history, foreclosure data, AVM | ATTOM API (paid) |
| HouseCanary | Property-level AVM, market analytics, risk scores | HouseCanary API (paid) |
| Freddie Mac PMMS | Weekly primary mortgage market survey | Public Excel download |
| Apartments.com | Active rental listings for rental comp analysis | Crawl4AI |
| Rentometer | Rental comp benchmark | Rentometer API |
Transit / Commute
| Source | What | How |
|---|---|---|
| NJ Transit GTFS | Train/bus static schedules (rail lines: NEC, NJCL, Montclair-Boonton, Morris-Essex, Raritan Valley, etc.) | Public GTFS feed |
| NJ Transit Real-Time | Live train delays, cancellations | NJ Transit Vision API |
| Google Maps Platform | Driving times, traffic, transit routing | Maps API |
| PATH (Port Authority) | Hudson County to NYC subway times | GTFS |
Communications & Calendar
| Source | What | How |
|---|---|---|
| Twilio | SMS showing confirmations, client property alerts | Twilio API |
| SendGrid | Email CMAs (PDF), market digest newsletters | SendGrid API |
| Google Calendar | Showing scheduling, availability check | Google Calendar API (OAuth2) |
Quick Start
Prerequisites
- Docker + Docker Compose
- Node.js 20+
- API keys (see
.env.example)
1. Clone and configure
git clone <repo>
cd nj-realtor-mcp
cp .env.example .env
# Edit .env with your API keys
2. Start the full stack
cd docker
docker compose up -d
This starts:
- PostgreSQL 16 with pgvector + PostGIS (geo queries + semantic search)
- Redis (caching + job queue)
- Crawl4AI (async browser-based crawler with anti-bot mode)
- MCP Server (the 20 tools)
- Cron service (background data sync)
3. Initialize the database
docker exec nj_realtor_db psql -U realtor -d nj_realtor -f /docker-entrypoint-initdb.d/01_schema.sql
node scripts/seed_nj_zips.js # seeds all NJ ZIP codes with county, transit, tax rate data
4. Seed MLS data
node scripts/seed_crawl.js # queues initial crawl jobs for NJ listings
5. Register with Claude Desktop
Add to ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"nj-realtor": {
"command": "node",
"args": ["/path/to/nj-realtor-mcp/src/index.js"],
"env": {
"POSTGRES_URL": "postgresql://realtor:pass@localhost:5432/nj_realtor",
"REDIS_URL": "redis://localhost:6379",
"CRAWL4AI_URL": "http://localhost:11235"
}
}
}
}
Example Claude Conversations
Property Due Diligence
"Run deep research on 47 Oak Avenue, Rumson NJ 07760"
Claude will: pull price history, run CMA, check permits/liens, assess flood risk, analyze commute to NYC, pull school district data, check for rent control (not applicable for SFR), and synthesize into a full due diligence report.
Buyer Matching
"I have a new buyer, Sarah Chen, looking for a 3-4BR under $750k in Monmouth County with good schools and under 60 min train to NYC. Set up alerts."
Claude will: save buyer profile, search active listings, rank by match score, send initial SMS with top 5 matches, and enable ongoing alerts for new listings.
Investment Analysis
"Analyze 22 Main St, Asbury Park NJ as a rental investment. Purchase price $620k."
Claude will: estimate market rent (Asbury Park has strong rental demand + rent control), run full NOI/cap rate/cash-on-cash analysis, flag NJ's high property tax rate, check for rent control, and calculate 5-year IRR.
Market Report
"Give me a market pulse for Monmouth County for the last 30 days"
Claude will: pull median prices, DOM, inventory, list-to-sold ratios, YoY changes, and write a 3-sentence market narrative for your newsletter.
Listing Copy
"Draft a listing for 8 Harbor View Dr, Sea Bright NJ. Highlights: oceanfront, fully renovated, elevator, 4BR, $3.2M."
Claude will: generate MLS description (500 words), Instagram caption, Facebook post, and 5 key feature bullets — all optimized for waterfront luxury segment.
NJ-Specific Intelligence
This MCP was built with deep NJ market knowledge baked in:
- Property taxes: NJ has the highest effective rate in the US (avg 2.23%). Every valuation and investment analysis explicitly accounts for this.
- Flood risk: Post-Hurricane Sandy, flood zone classification is critical for NJ coastal/shore properties (Ocean, Monmouth, Atlantic counties especially). We integrate FEMA + NJ DEP + sea level rise projections.
- School districts: NJ's #1 buyer decision factor. We integrate NJ DOE District Factor Groups (A through J), NJSLA scores, and GreatSchools ratings.
- NJ Transit: Door-to-door commute time to NYC Penn Station is a core value driver. We compute actual train journey times for 163 NJ rail stations.
- Rent control: ~100 NJ municipalities have rent control. Every investment analysis flags this automatically.
- 565 municipalities: Each has its own tax rate, zoning rules, and permit system. Our tax lookup normalizes across all of them.
- Mansion tax: NJ imposes a 1% mansion tax on residential sales over $1M. Our mortgage and investment calculators include this.
- Landlord-tenant law: NJ is tenant-friendly (Anti-Eviction Act). Investment analyses note this risk.
Architecture
Claude Desktop
│
▼ MCP (stdio)
NJ Realtor MCP Server (Node.js)
│
├── PostgreSQL 16 + pgvector + PostGIS
│ ├── properties table (MLS + ATTOM merged)
│ ├── sales_history (deed records)
│ ├── tax_assessments (NJ MOD-IV)
│ ├── permits_liens (NJ Courts + municipal)
│ ├── market_stats (pre-aggregated)
│ ├── clients (buyer profiles + alert criteria)
│ └── crawl_jobs (queue for background crawling)
│
├── Redis (6hr TTL cache for CMAs, neighborhood reports)
│
└── Crawl4AI (Docker, Playwright-based)
├── Zillow property pages
├── NJ county tax/deed portals (20 counties)
├── NJ Courts lis pendens search
├── NJ Flood Mapper
├── Apartments.com rental listings
├── NJ Transit schedule pages
└── SpotCrime crime data
Roadmap
- [ ] Matterport 3D tour integration (auto-generate for listed properties)
- [ ] NJ short-term rental ordinance tracker (Airbnb viability by municipality)
- [ ] HOA document analyzer (AI reads CC&Rs and flags red flags)
- [ ] Contractor permit history (who pulled permits, licensed/unlicensed work)
- [ ] Solar potential analysis (NJ has strong solar incentives — SREC program)
- [ ] 1031 exchange tracker (identify replacement properties meeting exchange criteria)
- [ ] MLS listing auto-publish (post approved listings directly to GSMLS)
- [ ] Offer analysis tool (compare multiple offers with net-to-seller calculation)
- [ ] Property condition report parser (AI reads inspection reports, extracts issues)
- [ ] NJ ANCHOR benefit calculator (NJ property tax relief program for homeowners)
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。