RE Data Refinery MCP¶
Pay-per-query real estate intelligence: live Zillow listings enriched with county-level data and scored for investment. RE Data Refinery turns messy property data into clean, scored, AI-ready intelligence for Columbus, OH and 14 surrounding metro cities — combining ZillAPI listings with GIS zoning, tax delinquency, sheriff sales, permits, and probate records, then computing flip score, wholesale score, rental yield, and market heat for every property.
Server type: Local (Python, stdio/SSE) + hosted worker
Auth: EVM wallet with USDC on Base (x402 pay-per-call)
Worker URL: https://re-data-refinery.ares-hms.workers.dev
Tools: 10
Pricing: $0.25-$0.50 USDC per lookup (free health/credits tools)
Category: Real Estate
Built by: RE Data Refinery (Columbus, OH)
Why This Matters for Operators¶
Real estate investors stitch together MLS feeds, county records, and their own scoring spreadsheets — usually across three or four subscriptions. RE Data Refinery collapses that into one agent-callable surface with per-query micropayments: agents pay $0.25–$0.50 USDC per lookup via x402 on Base, with no subscription, no API tiers, and no minimum commitment.
The differentiator is enrichment depth: county GIS, tax delinquency, sheriff sales, permits, and probate records sit alongside standard listing fields, and proprietary flip/wholesale/rental-yield/market-heat scores come pre-computed per listing. A free local fallback mode supports development without spending USDC.
Tools & Capabilities¶
| Tool | What it does | Price |
|---|---|---|
refinery_health |
API health, cached property count, rate limits | Free |
refinery_credits |
Upstream ZillAPI credit balance | Free |
refinery_properties |
Scored property lists by city with price filters | $0.35 (free cached) |
refinery_property_detail |
Full scored detail for one property (ZPID) | $0.35 |
refinery_property_price_history |
Price/transaction timeline | $0.25 |
refinery_property_tax_history |
Tax and assessment history | $0.25 |
refinery_property_schools |
School ratings near a property | $0.25 |
refinery_search |
Natural-language property search | $0.50 |
refinery_scored_search |
Search filtered by investment criteria | $0.50 |
refinery_payment_status |
x402 configuration and active API base | Free |
Installation¶
pip install "x402>=2.20.0" eth-account httpx
git clone https://github.com/areshms/re-refinery-mcp.git
cd re-refinery-mcp
export EVM_PRIVATE_KEY=0x...
Configuration¶
{
"mcpServers": {
"re_refinery_mcp": {
"command": "python3",
"args": ["/path/to/re-refinery-mcp/re_refinery_mcp.py"],
"env": {
"EVM_PRIVATE_KEY": "${EVM_PRIVATE_KEY}",
"X402_SPEND_CAP": "$1"
}
}
}
}
The wallet must hold USDC on Base mainnet for paid lookups. Payments settle automatically via the x402 protocol using Permit2 signatures and the Coinbase CDP facilitator. X402_SPEND_CAP sets a per-payment cap (default $1). MIT licensed.
Business Relevance¶
- Flippers and wholesalers source and score deals from chat instead of spreadsheets
- Rental investors filter by rental yield and market heat across 14 metro cities
- Analysts pull tax, price, and school histories per property on demand
- Agent builders fold property scoring into larger investment workflows with spend caps
Integration with CorpusIQ¶
RE Data Refinery covers property sourcing and scoring — a domain CorpusIQ's connectors do not touch. Paired in one agent session, an investor can source and score properties through RE Data Refinery while CorpusIQ handles the financial layer: QuickBooks for the business books, Stripe for payments, and banking for deal funding — then join the two on property address or LLC name. CorpusIQ's x402-aware spend discipline (caps, per-call pricing) mirrors RE Data Refinery's pay-per-query model.
Limitations¶
- Coverage is Columbus, OH metro (14 cities, 150+ seeded properties) — not national
- Requires a funded wallet and x402 client; USDC on Base only
- New listing (Aug 2026), zero-star repository, single maintainer
- Paid lookups depend on the hosted Cloudflare worker's uptime
- Local free mode uses a different (limited) data path than paid worker lookups