Zillow Mcp

zillow-mcp

[CI] [npm] [license]

Zillow real-estate access as an MCP server for Claude — search listings, fetch property details, Zestimate history, your saved searches & homes, and market reports via natural language.

⚠️ Zillow does not publish a public consumer API. The official Bridge API is gated to MLS partners. This server uses the same private endpoints the zillow.com web app uses, routed through your own signed-in browser tab via the ContextMint Bridge browser extension. Every request acts on behalf of your existing session — your cookies, your TLS, your JS context — exactly as if you'd clicked it in the browser yourself. Treat this as informal use of Zillow's website. Use at your own discretion.

Why this exists

The four existing Zillow MCPs all sit on one of two foundations:

• The Bridge API — requires MLS membership, IDX vendor relationship, or "approved technology partnership" (10+ business-day approval). Consumers can't get in.
• A paid scraper (RapidAPI, Apify) — adds a third party to the trust path and rate-limits.

None of them can see what you have saved, favorited, or recently viewed — because both Bridge and third-party scrapers are out-of-session. zillow-mcp uses your already-signed-in zillow.com tab.

Tools

Input and output contract

Changes to tool inputs and outputs, so you can adjust any scripts or prompts that call these tools:

• zillow_compare_properties reports bot-wall rows like zillow_bulk_get does. A row that Zillow keeps blocking with a captcha now comes back as error_kind: "bot_challenge" instead of a generic error kind, so you can tell it apart from a missing listing. The response can also include blocked (how many rows stayed blocked), retry_after_s (how many seconds to wait before retrying just those rows) and pending (how many rows did not finish before the overall deadline; those rows have error_kind: "pending").

• A zpid must be a number. Every tool that takes a zpid (zillow_get_property, zillow_bulk_get, zillow_compare_properties, zillow_get_property_photos, zillow_get_zestimate_history, zillow_get_price_history, zillow_get_tax_history) accepts a positive whole number, or a string of digits only such as "12345". Any other string, such as "12345_zpid" or a URL, is now rejected as invalid input. Pass a homedetails URL in the url or urls field instead.

• Pass zpids or urls, not both. zillow_bulk_get and zillow_compare_properties now return an error when a call includes both lists. Before, the urls list was silently ignored. To fetch both kinds, make two calls.

Upgrading

next release

• zillow_sweep_area on a hosted deployment (mcp-host, where the server and the caller share no filesystem). Running locally, nothing changes.
  • With output_path, the file is no longer written to the server's disk, which you could not reach. It goes to the host's file store as a zip, and the reply carries a download link. The reply has output_url (the signed link to the zip), output_expires_at (when the link stops working, 1 hour by default) and output_file (the file's name inside the zip, the base name of your output_path). It has no output_path.
  • If the host has no file store, an output_path call is refused before any request is made. Leave output_path out to get the listings inline.
  • Without output_path, the listings still come back inline. If they are larger than 10 MiB, they come back as a link instead (the same three fields plus a note, with no results), because the host would reject a reply that large. With no file store the call fails with a "too large" error that tells you to narrow the area.

Acknowledgement of Terms

By using this MCP server, you acknowledge and agree to the following:

1. This server accesses your own Zillow session. Every request is dispatched through your own browser tab (logged in or not) via the ContextMint Bridge extension. It does not — and cannot — access anyone else's account.

2. Zillow's Terms of Use govern your use of this server, just as they govern your direct use of zillow.com. The clauses most relevant here:

You may not use any robot, spider, scraper or other automated means to access the Services for any purpose without our express written permission… nor may you conduct automated queries (including screen and database scraping, spiders, robots, crawlers, bypassing CAPTCHAs or similar precautions).

You are agreeing to those terms — read by the maintainer 2026-05-23 — every time you invoke a tool in this server. Zillow's terms broadly prohibit automated access without written permission; this is an unofficial tool and Zillow has not granted it permission.

3. Personal, non-commercial use only. This project is not affiliated with, endorsed by, sponsored by, or in partnership with Zillow Group. It is a personal automation tool that drives the same Zillow website you would drive by hand — one search at a time, your own saved homes, your own market reports. Do not use it to bulk-extract listings, train models, populate a competing real-estate product, or for any commercial purpose.

4. Stability is not guaranteed. This server reads private internal endpoints (/async-create-search-page-state/, __NEXT_DATA__ blobs, /myzillow/...) that Zillow may change without notice. It may break. It may stop working. That's by design — the surface is not theirs to maintain on our behalf.

5. You accept full responsibility for any consequences of using this server in connection with your Zillow access — rate limiting, account warnings, suspension, IP blocks, captcha walls, or any enforcement action Zillow Group takes. If Zillow objects to your use, stop using this server.

This section is the maintainer's good-faith summary of the terms — it is not legal advice and does not modify or supersede Zillow's actual ToS.

Install

Option A — npx (after publishing)

Add to .mcp.json:

{  "mcpServers": {    "zillow": {      "command": "npx",      "args": ["-y", "zillow-mcp"]    }  }}

Option B — from source

git clone https://github.com/chrischall/zillow-mcpcd zillow-mcpnpm installnpm run build

{  "mcpServers": {    "zillow": {      "command": "node",      "args": ["/path/to/zillow-mcp/dist/bundle.js"]    }  }}

One-time browser setup

zillow-mcp talks to your browser through the ContextMint Bridge extension, which is shared across every fetchproxy-based MCP (resy-mcp, opentable-mcp, …). Install it once from the ContextMint Bridge releases:

• Chrome: download the Chrome zip, unzip it, then chrome://extensions → toggle Developer mode → Load unpacked → pick the unzipped folder.
• Safari: not available yet — it will ship inside the ContextMint app, which has no public download. Use Chrome for now.

ContextMint Bridge is the fetchproxy browser extension under its new name, from the same maintainer — fetchproxy's own README (Install → Extension) points to it. Its source is public at nullnet-app/contextmint-bridge: build it yourself (see that repo's README), or check a release zip against the .sha256 file published beside it before loading it — shasum -a 256 -c contextmint-bridge-chrome-<version>.zip.sha256.

Open zillow.com and sign in. That's all the auth this server needs.

How it works

┌────────────────┐  stdio   ┌──────────────────┐   WS   ┌──────────────────┐    fetch()    ┌─────────────┐│ MCP client     │◀────────▶│  dist/bundle.js  │◀──────▶│  ContextMint     │◀────────────▶│ zillow.com  ││ (Claude, etc.) │          │  (Zillow MCP)    │ :37149 │  Bridge          │   (real TLS, │ (your tab)  │└────────────────┘          └──────────────────┘        │  (separate)      │   cookies)    └─────────────┘

The MCP server runs in Node, but every HTTP call to zillow.com is dispatched into your live browser tab through the ContextMint Bridge extension. Each request rides your existing session — _abck, TLS fingerprint, and cookies all match the page that's already on screen. No headless browser stand-in, no separate identity, no third-party proxy: just your real browser, acting on its own behalf, with the MCP server picking what to ask for.

Commands

npm test               # vitest, mocked transport, no networknpm run test:watchnpm run test:coveragenpm run build          # tsc --noEmit + esbuild bundle → dist/bundle.jsnpm run dev            # node dist/bundle.js (after build)

License

MIT