mcp-nextdoor
An MCP server for Nextdoor that automates neighborhood feed browsing, marketplace search, event discovery, and local service finding.
README
@striderlabs/mcp-nextdoor
MCP (Model Context Protocol) server for Nextdoor — the local community platform. Automate neighborhood feed browsing, marketplace search, event discovery, and local service finding using Playwright with Browserbase.
Requirements
- Node.js 18+
- A Browserbase account with a CDP-enabled session URL
- A Nextdoor account
Environment Variables
| Variable | Required | Description |
|---|---|---|
BROWSERBASE_CDP_URL |
Yes | Browserbase CDP WebSocket URL for browser connection |
NEXTDOOR_EMAIL |
Yes | Your Nextdoor account email |
NEXTDOOR_PASSWORD |
Yes | Your Nextdoor account password |
Installation
npm install @striderlabs/mcp-nextdoor
Or from tarball:
npm install ./striderlabs-mcp-nextdoor-1.0.0.tgz
MCP Configuration
Add to your MCP client config (e.g., Claude Desktop ~/.claude/claude_desktop_config.json):
{
"mcpServers": {
"nextdoor": {
"command": "node",
"args": ["/path/to/node_modules/@striderlabs/mcp-nextdoor/dist/index.js"],
"env": {
"BROWSERBASE_CDP_URL": "wss://connect.browserbase.com?apiKey=YOUR_KEY&sessionId=YOUR_SESSION",
"NEXTDOOR_EMAIL": "you@example.com",
"NEXTDOOR_PASSWORD": "yourpassword"
}
}
}
}
Tools
get_feed
Get posts from your Nextdoor neighborhood feed.
Parameters:
limit(number, 1-50, default: 10) — Number of posts to retrievecategory(string, default: "all") — Filter by:all,crime_safety,lost_found,free_stuff,recommendations,general
Example:
{ "limit": 5, "category": "crime_safety" }
create_post
Draft a new post in your Nextdoor neighborhood. The post is loaded into the composer but not auto-submitted (safety measure).
Parameters:
title(string, required) — Post titlebody(string, required) — Post body contentcategory(string, default: "general") — Category:general,crime_safety,lost_found,free_stuff,recommendations
Example:
{ "title": "Lost black cat - Maple St area", "body": "Missing since Tuesday evening...", "category": "lost_found" }
search_marketplace
Search Nextdoor's For Sale & Free marketplace.
Parameters:
query(string, required) — Search termsmax_price(number, optional) — Maximum pricemin_price(number, optional) — Minimum pricelimit(number, 1-50, default: 10) — Number of listings
Example:
{ "query": "bicycle", "max_price": 200, "limit": 5 }
get_events
Get upcoming local events from Nextdoor.
Parameters:
limit(number, 1-50, default: 10) — Number of eventsdays_ahead(number, 1-90, default: 30) — Days ahead to look
Example:
{ "limit": 10, "days_ahead": 14 }
get_recommendations
Get local business recommendations from neighbors.
Parameters:
category(string, default: "all") — Business type (e.g.,restaurant,plumber,dentist)limit(number, 1-50, default: 10) — Number of recommendations
Example:
{ "category": "restaurant", "limit": 5 }
find_services
Search for local service providers.
Parameters:
service_type(string, required) — Service to find (e.g.,plumber,electrician,babysitter)limit(number, 1-50, default: 10) — Number of providers
Example:
{ "service_type": "house cleaner", "limit": 5 }
How It Works
The server uses Playwright to connect to a Browserbase-managed Chromium browser via CDP (Chrome DevTools Protocol). It navigates to Nextdoor, authenticates with your credentials, and scrapes or interacts with the page to fulfill tool requests.
MCP Client → mcp-nextdoor server → Playwright CDP → Browserbase → Chromium → nextdoor.com
Notes
- Authentication: The server logs in using email/password on first use. Sessions are reused across tool calls.
- Post creation:
create_postdrafts content into the composer but does not submit automatically to prevent accidental posts. - Selectors: Nextdoor's frontend changes periodically. If extraction returns empty results, the page structure may have changed.
- Rate limiting: Be respectful of Nextdoor's terms of service and avoid excessive automated requests.
License
MIT — Strider Labs
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。