Housing Intel MCP — Meta-pack that chains FRED, BLS, ATTOM, and HUD APIs
One address for any MCP client. No sign-in needed on this server.
At most: only reads or looks up information.
Odel charges nothing per call. The publisher may charge for its own service.
Odel could not list its tools without signing in.
Housing Intel MCP — Meta-pack that chains FRED, BLS, ATTOM, and HUD APIs
Part of Pipeworx — an MCP gateway connecting AI agents to 1683+ live data sources.
| Tool | Description |
|---|---|
case_shiller_metro_compare | Compare Case-Shiller home price indices across multiple US metros in one call (the 20-city composite). For each metro returns latest level, 3-month change, 12-month change, all-time peak, drawdown from peak, and a softening flag. Output also ranks metros softest → strongest. Use for "which metros are softening", "Case-Shiller for [list of cities]", "compare housing prices in X, Y, Z" queries — picks the right per-metro FRED series IDs (DNXRSA, PHXRSA, TPXRSA, etc.) so callers don't have to. Available metros: Atlanta, Boston, Charlotte, Chicago, Cleveland, Dallas, Denver, Detroit, Las Vegas, Los Angeles, Miami, Minneapolis, New York, Phoenix, Portland, San Diego, San Francisco, Seattle, Tampa, Washington DC. NOT A CURRENT-MARKET READ: Case-Shiller is monthly, published on a roughly two-month lag, and each value is a three-month moving average — so the newest observation is typically 90-120 days old and describes a trailing quarter. The freshness block states the exact age of the call you made. |
housing_market_snapshot | Get national housing market overview: mortgage rates, housing starts, Case-Shiller index, unemployment, construction employment. Optionally add metro-level prices (e.g., "Denver", "Atlanta"). For comparing Case-Shiller across multiple metros use case_shiller_metro_compare instead. THE COMPONENTS DO NOT SHARE A DATE — they are weekly, monthly and quarterly official statistics, each published on its own lag, so one call can span months of vintages. freshness states the span in days and names the oldest component; read it before comparing components against each other. |
housing_property_report | Analyze a property by address and zip code. Returns valuation estimate, sales history, tax assessment, and detailed characteristics. |
housing_rental_analysis | Evaluate rental investment potential by address and zip code. Returns estimated rent, fair market rents, and CPI rent trends. |
housing_affordability_check | Check housing affordability in a market. Returns mortgage rate, median price, monthly payment, required income, and HUD limits. Optionally specify metro (e.g., "Denver"). The inputs are official statistics on DIFFERENT cadences — a weekly mortgage rate against a quarterly median price — so the monthly payment is an illustrative ratio of published figures, not a live quote. freshness states the span and the age of each input. |
housing_employment_outlook | Assess labor market health for housing demand. Returns employment, construction jobs, residential building employment, unemployment rate, and job openings. All BLS monthly releases; each series carries its own observation_date and cadence (JOLTS runs an extra month behind CES/LAUS), and freshness states the span. snapshot_date is the call date, not the data date. |
housing_signal_scan | Scan 45+ housing indicators for anomalies and reversals. Flags unusual moves across rates, starts, sales, prices, wages, unemployment, and rent. Each signal carries its own observation_date and cadence — series here span weekly (mortgage rate) to quarterly (median sales price), and Case-Shiller is additionally a 3-month moving average, so "turned negative" does not mean "this month" for every series. freshness states the full span and the oldest component. |
housing_mortgage_history | Freddie Mac Primary Mortgage Market Survey — the weekly US mortgage INTEREST rate (the annual percentage borrowers pay on a home loan, e.g. 6.5%), back to 1971. This is the borrowing cost paid by home buyers. Returns the latest snapshot, a time series for the requested window, and min/max/avg stats. Sourced from Freddie Mac directly (not FRED), ingested weekly by the Pipeworx data pipeline. |
housing_market_screen | Rank US metros for rental cash flow in ONE call — the "which markets are best for a landlord" view. Returns metros sorted by gross rent yield = (Zillow median monthly rent × 12) ÷ Zillow typical home value. No per-metro orchestration and no API key. Use for "best/worst rental markets", "highest-yield metros", "where does rent go furthest vs. home prices". Tune with direction (top = highest yield / best cash flow, bottom = lowest), limit, and optional home-value bounds. rent_as_of and home_value_as_of are two independent Zillow queries and can land on different months; freshness says whether they matched this call. |
housing_metro_demand | Demand + rent-durability signals for a shortlist of US metros in ONE call — population & 5-year growth, renter share, median household income, and unemployment, straight from Census ACS. Deterministic by metro (CBSA-keyed) — NO FRED series-ID guessing. Pass metros ("City, ST", e.g. the top results from housing_market_screen). This is the Stage-2 "is the demand real?" filter on a yield shortlist — high yield in a shrinking metro is a trap. No API key needed. HIGHEST-LAG SOURCE IN THIS PACK: figures are a Census ACS 5-YEAR rolling average (currently 2018-2022, released Dec 2023) — for structural demand, not a current-quarter read. freshness states the exact age. |
Add to your MCP client (Claude Desktop, Cursor, Windsurf, etc.):
{
"mcpServers": {
"housing-intel": {
"url": "https://gateway.pipeworx.io/housing-intel/mcp"
}
}
}
tools/list at https://gateway.pipeworx.io/housing-intel/mcp returns the tools in the table
above plus the shared Pipeworx meta-tools — ask_pipeworx,
discover_tools, search_within, remember/recall and the rest of the
gateway-wide set. So the tool count you see is larger than this table: a
single-pack endpoint currently lists roughly 30 shared tools alongside the
pack's own. The connection's initialize response states its exact scope, and
is the authoritative answer for a given day.
This is deliberate, not multiplexing by accident. The meta-tools are what let a
scoped connection answer a question this pack does not cover — via
ask_pipeworx, which routes across the whole catalog — without you adding a
second MCP server. There is currently no way to mount a pack endpoint without
them; if the extra schemas cost you more context than the routing is worth,
connect to the full gateway once rather than to several pack endpoints.
Or connect to the full Pipeworx gateway to get every pack's tools listed directly, instead of just this one's:
{
"mcpServers": {
"pipeworx": {
"url": "https://gateway.pipeworx.io/mcp"
}
}
}
Both URLs reach the same gateway and the same 1683+ data sources. The
only difference is which pack's tools are listed directly; ask_pipeworx
reaches all of them from either one.
curl -X POST https://gateway.pipeworx.io/v1/tools/case_shiller_metro_compare \
-H 'Content-Type: application/json' \
-d '{"metros":["Denver","Phoenix","Tampa","Charlotte"]}'
No account needed for the first calls. Inspect any tool: GET https://gateway.pipeworx.io/v1/tools/case_shiller_metro_compare. Find one: POST https://gateway.pipeworx.io/v1/tools/search_packs with {"query":"..."}.
This package also runs as a local stdio MCP server — no Pipeworx account, no gateway round-trip:
{
"mcpServers": {
"housing-intel": {
"command": "npx",
"args": ["-y", "@pipeworx/mcp-housing-intel"]
}
}
}
Or run it directly to confirm it starts:
npx -y @pipeworx/mcp-housing-intel
It speaks MCP over stdin/stdout and answers initialize/tools/list/tools/call
for only this pack's tools — none of the shared meta-tools the gateway
connection above adds. Same source, same tools, no ask_pipeworx routing.
Instead of calling tools directly, you can ask questions in plain English — this works on the pack endpoint above as well as on the full gateway:
ask_pipeworx({ question: "your question about Housing Intel data" })
The gateway picks the right tool and fills the arguments automatically.
MIT