Hud MCP Server

HUD — Housing and Urban Development

The U.S. Department of Housing and Urban Development's data. Fair Market Rents (FMR), Income Limits, USPS Vacancy Data, public housing characteristics, Continuum of Care (homelessness data), Comprehensive Housing Affordability Strategy (CHAS) data. Free, no auth.

Part of Pipeworx — an MCP gateway connecting AI agents to 1476+ live data sources.

Why this matters for AI agents

For housing-affordability analysis, federal benefit thresholds, or homeless services research, HUD is the source. Pair with Census (housing characteristics), Altos (live market data), ATTOM (property-level data), and FRED (mortgage/macro context).

Common flows:

• Fair Market Rent. "What's HUD's FMR for a 2-bedroom in Denver?" — FMR data drives Section 8 voucher amounts.
• Income Limits. "Median income for a family of 4 in Boston?" — drives eligibility for housing programs.
• USPS Vacancy. "What % of housing units are vacant in this ZIP?" — quarterly USPS-derived vacancy rates by tract.
• Homelessness. Point-in-Time count, demographic breakdowns, Continuum of Care reports.
• Public Housing. Authority listings, unit counts, demographic served.

Auth

Most HUD APIs are free, no auth. A few require registering for a free token. Pipeworx supports _apiKey passthrough where applicable.

Datasets worth knowing

Common pitfalls

• FMR vs. market rent. Fair Market Rent is HUD's policy number used for voucher reimbursement. It's NOT what the market actually charges — typically lags real market rents by 6-18 months. For market reality, use Altos or rental aggregator data.
• AMI varies by household size. "80% of AMI" depends on whether the household has 1, 2, 3, 4+ people. The "Median Income" headline doesn't capture this.
• Vacancy rates from USPS. USPS vacancy isn't true vacancy — it counts units the postal carrier marks as vacant for 90+ days. May undercount short-term vacancies and overcount foreclosures.
• Homeless point-in-time counts undercount. Most jurisdictions do their PIT count one night in late January. Methodology varies by Continuum of Care, and unsheltered counts especially miss people. Treat as a lower bound.
• CHAS data is 5-year smoothed. Like ACS 5-year data — current as of 5 years rolling. For real-time housing affordability changes, layer in faster signals.
• CBSA vs. county FMRs. HUD publishes FMRs at the CBSA (metropolitan/micropolitan area) level. Specific counties may have very different actual rents from their CBSA's FMR average.

Quick Start

Add to your MCP client (Claude Desktop, Cursor, Windsurf, etc.):

{
  "mcpServers": {
    "hud": {
      "url": "https://gateway.pipeworx.io/hud/mcp"
    }
  }
}

What this endpoint actually serves

tools/list at https://gateway.pipeworx.io/hud/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 1476+ data sources. The only difference is which pack's tools are listed directly; ask_pipeworx reaches all of them from either one.

Using with ask_pipeworx

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 Hud data" })

The gateway picks the right tool and fills the arguments automatically.

More

• Docs and guides
• pipeworx.io

License

MIT