angi-mcp

angi-mcp

MCP server for Angi (Angie's List) that lets you search for home-service pros by trade and city, read ratings, profiles, and reviews, and manage your Angi projects using an existing browser session via fetchproxy.

Category
访问服务器

README

angi-mcp

MCP server for Angi (formerly Angie's List) — find home-service pros by trade and city, and read their ratings, profiles and reviews.

Angi serves its pages only to a real browser, so requests route through the user's own angi.com tab via the fetchproxy browser extension, reusing their existing session. The trade/city taxonomy is read directly from Angi's public sitemaps and needs no browser at all.

No Angi account or credentials are required. Everything this server reads is public.

This project was developed and is maintained by AI (Claude). Use at your own discretion.

Install

npm install -g @chrischall/angi-mcp

Register it with your MCP host:

{
  "mcpServers": {
    "angi": { "command": "angi-mcp" }
  }
}

You also need the fetchproxy Transporter browser extension, with an open angi.com tab and its site access allowing angi.com. On the first request the extension shows a pairing code to approve; the trust then persists.

Run angi_healthcheck to confirm the bridge is connected.

Tools

Tool What it does
angi_search_pros Pros for a trade in a US city, with ratings, review counts, years in business, service area, amenities. 10 per page.
angi_get_pro One pro's full profile and ratings breakdown.
angi_get_reviews Reviews on a pro's profile — rating, text, reported cost, date, categories, the pro's response. Filterable by rating.
angi_list_trades Every trade slug Angi publishes (~312). No bridge needed.
angi_list_cities Cities Angi publishes pages for, per trade. No bridge needed.
angi_healthcheck Bridge connection state.

Signed-in tools (need the browser tab signed in to Angi):

Tool What it does
angi_get_account Your identity and open/closed project counts.
angi_list_my_projects Your Angi projects, open and closed.
angi_list_my_reviews Reviews you've written, plus pros awaiting a rating.

Searches take slugs, not free text — resolve them first:

angi_list_trades   { contains: "duct" }      -> "air-duct-cleaning"
angi_list_cities   { trade: "plumbing", state: "nc", contains: "char" }
angi_search_pros   { trade: "plumbing", state: "nc", city: "charlotte", compact: true }

Pass compact: true when browsing or ranking — it projects each record to a slim summary instead of the full ~1KB payload.

Ratings

Each pro carries two overall ratings and they differ on purpose:

  • rating (averageRatings.OVERALL) — unrounded. Rank on this.
  • displayRating (averageOverallRating) — the rounded value Angi shows.

Per-dimension ratings (quality, value, punctuality, professionalism, responsiveness) come through in the full record.

isSponsored: true marks paid placement, not a quality signal.

Limits

  • No zip-code filtering. Angi's ?zip= parameter is inert — a Charlotte URL returns Charlotte pros regardless. Location comes from the city slug.
  • Account record shapes are unverified. The account used to map my.angi.com held zero projects and zero reviews, so only the envelopes were observed. angi_list_my_projects and angi_list_my_reviews return records raw and set recordFieldsVerified: false rather than projecting onto field names nobody has seen. See docs/ANGI-API.md.
  • The inbox is not readable. Angi's messages run on the Twilio Conversations SDK; the proxy endpoints reject the session cookie alone (401), so messages need the Twilio client rather than an HTTP GET.
  • Cannot be hosted remotely. Like every browser-bridge server in this fleet, it needs a signed-in tab on the machine it runs on, so it cannot be served to claude.ai from mcp-host.

Without the MCP

skills/angi-fpx/ is a shell-out skill that reads the same data with the fpx CLI and no running server — useful in scripts or on a machine where the MCP isn't installed.

Development

npm install
npm run build      # tsc --noEmit + esbuild bundle
npm test

docs/ANGI-API.md records the live-captured request/response shapes, including the verified negative results. Read it before changing src/parse.ts.

License

MIT

推荐服务器

Baidu Map

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。

官方
精选
JavaScript
Playwright MCP Server

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。

官方
精选
TypeScript
Magic Component Platform (MCP)

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。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

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。

官方
精选
本地
TypeScript
VeyraX

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。

官方
精选
本地
graphlit-mcp-server

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。

官方
精选
TypeScript
Kagi MCP Server

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。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

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。

官方
精选