homedata-mcp
Enables AI assistants to query UK property data including EPCs, sale history, planning, flood risk, council tax, demographics, and more via the Homedata API.
README
Homedata MCP Server
A Model Context Protocol server that exposes the Homedata UK property data API as native tools for AI coding assistants - Claude Desktop, Claude Code, Cursor, and any other MCP client.
29M UK addresses keyed by UPRN, with EPCs, sale history, planning, flood risk, council tax, demographics, crime, schools, broadband and transport - all queryable directly from your assistant chat.
Data is sourced from Home.co.uk's 30-year panel of partners, the Environment Agency, ONS Census 2021, the Valuation Office Agency, Ofcom, Ofsted, data.police.uk and HM Land Registry.
Install
Requires Python 3.10+.
pip install homedata-mcp
For local development from this repository:
git clone https://github.com/wehomemove/homedata-mcp.git
cd homedata-mcp
pip install -e .
You will need a Homedata API key. Sign up at homedata.co.uk/developer - the Free tier gives 100 calls / month with no card required.
export HOMEDATA_API_KEY=hd_live_xxx
homedata-mcp --help
CLI
The package also installs a homedata command — same data as the MCP
server but for human shells, scripting, and CI. Useful for quick lookups,
demos, and piping into other tools.
homedata property 100021421083
homedata epc 100021421083 --field current_energy_efficiency
homedata search "10 downing street" --postcode SW1A2AA
homedata flood 100021421083 --compact
homedata batch 100021421083 100022121211
Pass --field <dotted.path> to extract a single value (handy for shell
pipelines) and --compact for single-line JSON. Reads the same
HOMEDATA_API_KEY env var as the MCP server. Run homedata --help for
the full list of subcommands.
Wire it into Claude Desktop
Edit ~/Library/Application Support/Claude/claude_desktop_config.json
(macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"homedata": {
"command": "homedata-mcp",
"env": {
"HOMEDATA_API_KEY": "your_api_key_here"
}
}
}
}
Restart Claude Desktop. You should see the 16 Homedata tools appear in the tool picker.
Cursor / Claude Code
Both Cursor and Claude Code accept the same MCP server spec. Add an entry
pointing command at homedata-mcp and put your key in env.
Tools
All UPRN tools take a single 12-digit Unique Property Reference Number; all postcode tools accept any common UK postcode format.
| # | Tool | Endpoint | Inputs |
|---|---|---|---|
| 1 | lookup_property |
GET /api/properties/{uprn}/ |
uprn |
| 2 | lookup_epc |
GET /api/epc-checker/{uprn}/ |
uprn |
| 3 | lookup_flood_risk |
GET /api/flood-risk/?uprn= |
uprn |
| 4 | lookup_council_tax |
in development — returns a "coming soon" response | uprn |
| 5 | search_property_listings |
GET /api/property_listings/?uprn= |
uprn |
| 6 | get_comparables |
GET /api/comparables/{uprn}/?count= |
uprn, count (default 20, max 200) |
| 7 | get_planning_applications |
GET /api/planning/search/?uprn= |
uprn |
| 8 | get_demographics |
GET /api/demographics/?postcode= |
postcode |
| 9 | get_crime |
GET /api/crime/?postcode= |
postcode, date (YYYY-MM, optional) |
| 10 | get_schools |
GET /api/schools/?uprn=&radius_m= |
uprn, radius_m (default 1000) |
| 11 | get_broadband |
GET /api/broadband/?postcode= |
postcode |
| 12 | get_transport |
GET /api/transport/?uprn=&radius_m= |
uprn, radius_m (default 800) |
| 13 | get_postcode_profile |
GET /api/postcode-profile/?postcode= |
postcode |
| 14 | search_address |
GET /api/address/find/?q= |
query, postcode (optional) |
| 15 | get_property_sales |
GET /api/property_sales/?uprn= |
uprn |
| 16 | batch_property_lookup |
POST /api/property/batch/ |
uprns (list, max 50) |
Typical workflow
- Resolve text → UPRN with
search_address. - Look up the property:
lookup_property,lookup_epc,lookup_council_tax,lookup_flood_risk. - Add market context:
search_property_listings,get_property_sales,get_comparables. - Add area context:
get_postcode_profile(cheap one-shot), or callget_demographics/get_crime/get_schools/get_broadband/get_transportindividually.
Configuration
| Variable | Required | Default | Notes |
|---|---|---|---|
HOMEDATA_API_KEY |
yes | - | Your Homedata API key (Authorization: Api-Key ...). |
HOMEDATA_BASE_URL |
no | https://api.homedata.co.uk |
Override for staging or self-hosted. |
Every tool returns either the parsed JSON body from the API, or
{"error": "...", "status_code": N, "detail": ...} on failure - tools never
raise exceptions out into the MCP protocol layer.
Pricing
Calls made via this MCP server count against your Homedata API plan exactly the same as any other API call. Plans (as of writing):
| Plan | Price | Calls / month |
|---|---|---|
| Free | £0 | 100 |
| Starter | £49 | 2,000 |
| Growth | £149 | 10,000 |
| Pro | £349 | 50,000 |
| Scale | £699 | 250,000 |
See homedata.co.uk/pricing for the current list.
Development
pip install -e ".[dev]"
HOMEDATA_API_KEY=hd_live_xxx pytest -v
Tests are skipped automatically when no API key is present, so the suite is safe to run in CI without secrets.
Project layout
homedata-mcp/
pyproject.toml
README.md
homedata_mcp/
__init__.py
server.py # FastMCP server + CLI entry point
client.py # httpx wrapper with auth + uniform error handling
tools/
__init__.py
property.py # lookup_property, batch_property_lookup
epc.py # lookup_epc
risk.py # lookup_flood_risk, lookup_council_tax
listings.py # search_property_listings, get_property_sales, get_comparables
planning.py # get_planning_applications
local.py # get_demographics, get_crime, get_schools, get_broadband, get_transport
address.py # search_address
profile.py # get_postcode_profile
tests/
test_tools.py
License
MIT - see LICENSE.
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。