Parcel-GIS-mcp
Enables querying public ArcGIS REST services for county/state parcel data, GIS layer identification, and USGS elevation without API keys.
README
parcel-gis-mcp-server
Third connector in the Hutton SIR Research Agent suite (alongside the FEMA NFHL and Municode connectors). Exposes county/state parcel data and GIS layer identify queries by hitting public government ArcGIS REST services directly — the same "bypass the JS portal, query the service underneath" pattern that made the FEMA connector work.
No API keys. Every upstream source is a public endpoint.
What it solves
The single most consistent failure class in SIR runs: Tier-1 parcel portals (qPublic/Schneider, pbcpao.gov, OCPA, LOJIC apps) are JS-rendered or token-blocked, and the only Tier-2 mirror (FloridaParcels) is FL-only and unreliable. This server resolves the SIR §0A primary key — the parcel identifier — against the open ArcGIS layers those portals sit on top of, and adds point-in-polygon lookups (zoning district, land use, WMD, city limits) plus USGS elevation.
Tools
| Tool | Purpose |
|---|---|
parcel_lookup |
Parcel by identifier (§0A primary key) → attributes + WGS84 centroid. Exact → LIKE → alias-field fallback. |
parcel_search_by_address |
§0A(5) fallback path when no identifier was supplied. |
gis_layer_identify |
Point-in-polygon against a registered layer or any ArcGIS layer URL — zoning district, land use, municipal boundary, WMD split checks. |
gis_elevation_at_point |
USGS EPQS/3DEP ground elevation (SIR topography path 1, Quality Rule 8). |
gis_raw_query |
Escape hatch (mirrors municode_raw_get): list services, read layer schemas, hand-built queries — how new jurisdictions get onboarded. |
gis_registry_list |
Show all registered sources/layers with verification status. |
The centroid from parcel_lookup chains directly into the FEMA connector
(fema_flood_zone_lookup) and gis_elevation_at_point — parcel ID in, flood
zone + elevation out, with coordinates carried per Quality Rule 8.
The registry (src/registry.ts)
Coverage grows by adding a registry entry, not writing code — the code-side
mirror of the Jurisdiction Portal Registry. Every entry carries a verified
date or false (= ⟨verify at run⟩). Shipped seed:
fl-statewide— FGIO/FDOR cadastral, all 67 FL counties (✅ verified 2026-08-17; schema confirmed: PARCEL_ID, OWN_NAME, PHY_ADDR1, LND_SQFOOT, DOR_UC, S_LEGAL, SALE_*, JV)ky-jefferson— LOJIC open OpenDataPVA parcels (✅ verified 2026-08-17 — the open server at gis.lojic.org, not the token-blocked apps.lojic.org one; geometry + PARCELID/PIN/LRSN only)ky-state-pva-pattern— kygisserver per-county PVA template (⟨verify at run⟩; Webster confirmed present, Jefferson/Scott confirmed absent)tn-statewide-tdec— TDEC statewide parcels mirror for the Chattanooga market (⟨verify at run⟩; returned 403 to one non-browser client 2026-08-17)- Identify layers:
ky-jefferson-landuse(✅ listed),fl-wmd-boundaries+fl-municipal-boundaries(URL to pin at first use)
Onboarding a new jurisdiction: gis_raw_query the county's
/rest/services directory → find the parcel/zoning layer → read its field
schema (f=json on the layer root) → add a ParcelSource/IdentifyLayer
entry → commit → Render auto-deploys.
Local development
npm install
npm run build
npm start # listens on :3000, MCP endpoint = POST /mcp
node test/live-test.mjs # smoke test against localhost
Note: from inside a sandboxed/proxied environment outbound calls may be blocked; run the smoke test against the deployed URL instead:
node test/live-test.mjs https://YOUR-SERVICE.onrender.com/mcp
Deploy (GitHub → Render)
Same process as the FEMA and Municode connectors — see DEPLOYMENT.md for the
full step-by-step. Short version: push this repo to GitHub → Render "New Web
Service" from the repo → build npm install && npm run build, start
npm start → connector URL is https://<service>.onrender.com/mcp.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。