Rental Market Data
I built a read-only connector that lets your AI assistant (Claude, ChatGPT or Codex) screen a U.S. site for build-to-rent and multifamily: what is permitted or entitled nearby, what comparable rentals are asking after concessions, who lives within a few miles, and how the ZIP’s home values compare with its metro. There is no account and nothing to pay for.
Add it to Claude or Codex
OAuth · one consent click · no accountThe button opens Claude’s custom-connector dialog with the name and URL filled in. To add it by hand: Settings → Connectors → Add custom connector, then paste the server URL.
Codex: the button copies the server URL and opens OpenAI’s Codex setup guide in a new tab. In Codex open Settings → MCP servers → Add server, choose Streamable HTTP, paste the URL and save, then select Authenticate. URL copied.
- Server URL
https://mcp.hermanchan.ai/mcp- Sign-in
- After you add it, your AI assistant opens a consent page. Approve it and clear a quick Cloudflare Turnstile check. That creates an anonymous connection, identified only by a random id. No email, no password, no profile.
- Access
- Read-only. All nine tools look data up; none of them write anything.
Example prompts
Paste one into a chat with the connector on-
“Screen 2200 Katy Fort Bend Rd, Katy TX for a build-to-rent site.”
One call to site_screen returns permitted and entitled supply by stage within 1, 3 and 5 miles, the market’s rent medians and the ten nearest comps, ring demographics, and the ZIP’s home value index against its metro.
-
“What’s the rent picture for 3-bed townhomes within 3 miles of 505 W Baseline Rd, Tempe AZ?”
demand_rings places the address, rent_comps_near returns the nearest 3-bed floor-plan groups with asking and net-effective rent, and rent_market_summary gives the Phoenix median to read them against.
-
“How much multifamily supply is permitted within 5 miles of 1100 Peachtree St, Atlanta, and what stage is it at?”
supply_near groups nearby projects by stage (entitlement case, stormwater notice, site plan, issued permit, built) and names the nearest ones.
-
“Which builders have the most late-stage townhome projects in the Nashville metro?”
supply_search searches the same pipeline by name, metro, stage and size instead of by location.
-
“Is there any rent coverage in Huntsville, and how fresh is it?”
data_coverage answers from the data itself: which metros hold rows, every rent market with its collection date, and the fields the connector leaves out.
The data, and where it covers
Four layers · nine toolsSupply pipeline
Residential projects by stage, from entitlement cases and stormwater notices to issued permits, across nine metros: Atlanta, Austin, Charlotte, Houston, Huntsville, Las Vegas, Nashville, Phoenix and Tampa. Outside them the supply layer says so instead of returning an empty answer.
Rents
Asking and net-effective rents with concessions, collected weekly from operators’ own websites in 22 rent markets. Every row carries its collection date and the offer as the operator wrote it.
Demographics
Population, households, income, jobs and labor force in rings around any U.S. address, from the Census Bureau (ACS 5-year, 2020 Census, LEHD LODES) and the Bureau of Labor Statistics. The nine metros answer in about a second; elsewhere a call can take tens of seconds.
Home values
The Zillow Home Value Index for a ZIP against its metro, for any U.S. ZIP the index covers. It is an index level for the area, not an appraisal and not a price for any one property.
How to read the numbers
- Supply stages are views of the same market, not a funnel to add up. One project can appear as an entitlement case, then a stormwater notice, then a permit. Only issued permits not yet built are marked committed; a permit that is already built shows as “Built — delivered”, existing stock rather than pipeline.
- Early-stage sizes are estimates. Where a filing gives only acreage, the size is estimated from it and labelled that way. Entitlement cases sit one to four years ahead of construction, and many never build.
- Asking and net-effective rent are different numbers. Each comp carries both, plus the concession in weeks. A “starting at” price is flagged as a floor.
- A rent row is a floor-plan group (community, bed count and plan), not a whole community.
The nine tools
| Tool | What it answers |
|---|---|
site_screen | One address in, every layer out: supply by stage at 1, 3 and 5 miles, rents, demographics and home values, with a note wherever a layer has no data. |
demand_rings | Population, households, income, jobs and labor force in 1, 3, 5 and 10-mile rings, apportioned over 2020 census blocks. |
supply_near | Permitted and entitled residential supply within a radius, grouped by stage; only issued permits not yet built are marked committed, and built projects show as existing stock. |
supply_search | The same supply data searched by project or builder name, metro, stage or size. |
rent_comps_near | Nearest rental comps to a point, one row per floor-plan group, with asking range, net-effective range, concession weeks and the operator’s own wording. |
rent_market_summary | Median asking and net-effective rent for a whole market by bed count, with the denominators named. |
rent_community_lookup | Turns a community name into an address and coordinates. |
home_values | Zillow Home Value Index (ZHVI) level for a ZIP against its metro, with the ratio between them. |
data_coverage | What the connector can and cannot answer, read from the data. |
What is left out, and why
The licensed rental-community unit counts and the owner and property-manager names come from sources I can’t republish, so they are absent from everything the public service can read rather than filtered out afterwards. Unit counts and year built from public records are shown instead, each with its source and as-of date: county property records, Florida DBPR apartment licences and City of Austin permits. Where no public record covers a community its unit count is left out, and its year built falls back to third-party market data, labelled as such. Sale comps are not part of the connector at all: it sticks to public records, operators’ own listing pages and public statistics.
Limits
Free tier- Monthly
- 50 tool calls per connection per calendar month (UTC). Only successful calls count; invalid input and listing the tools are free.
- Remaining
- Every answer reports how many calls the connection has left this month.
- Pace
- Up to 10 calls a minute per connection, under a service-wide daily cap.
- Size
- Answers return 25 rows by default (15 for rent comps) and 50 at most, and say when a list was trimmed.
Sources and support
- Supply: public county and state records: building permits, subdivision and rezoning cases, plats, and construction stormwater notices filed with state environmental agencies.
- Rents: my own weekly collection from operators’ public listing pages, dated on every row.
- Demographics: U.S. Census Bureau (American Community Survey 5-year, 2020 Census PL 94-171, LEHD LODES, TIGER) and the U.S. Bureau of Labor Statistics.
- Home values: Zillow Home Value Index (ZHVI), Zillow Group, zillow.com/research/data. A static snapshot; when the source files carry no date, answers say the as-of date is unknown.
Neither Zillow, the Census Bureau, BLS nor any permitting agency endorses this connector.
- Support
- Questions, corrections or a request to switch a connection off: herman@hermanchan.ai. What to include and what to expect: support.
- Privacy
- What the service records and for how long: privacy policy.
- Terms
- The rules of use, in plain words: terms of service.