@striderlabs/mcp-angi

@striderlabs/mcp-angi

Angi (formerly Angie's List) home services connector for personal AI agents. Search for vetted service providers, request quotes, book appointments, track projects, and manage payments — all through natural language.

Built by Strider Labs.

Features

• Search providers — Find plumbers, electricians, HVAC techs, general contractors, cleaners, and more by location
• Provider profiles — View ratings, reviews, years in business, and contact info
• Request quotes — Submit quote requests for home projects
• Book appointments — Schedule service appointments with providers
• Message providers — Send and receive messages with service pros
• Project tracking — View active and completed home projects
• Payment management — Review payment history and invoices
• Booking history — See all past and upcoming appointments

Installation

npm install -g @striderlabs/mcp-angi

After installing, set up Playwright's browser binaries:

npx patchright install chromium

MCP Client Configuration

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

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

Generic MCP Client

{
  "mcpServers": {
    "angi": {
      "command": "npx",
      "args": ["-y", "@striderlabs/mcp-angi"]
    }
  }
}

Tool Reference

Tool	Description
angi_status	Check login status and session info
angi_login	Open Angi login page to authenticate
angi_logout	Log out and clear session data
angi_search_providers	Search for home service providers by type and location
angi_provider_profile	Get provider profile, ratings, and reviews
angi_request_quote	Request a quote for a home project
angi_book_appointment	Book an appointment with a provider
angi_get_messages	Retrieve messages from your Angi inbox
angi_send_message	Send a message to a provider
angi_get_projects	View active and completed home projects
angi_payment_history	View payment history and invoices
angi_booking_history	View past and upcoming appointments

Usage Examples

Find a plumber

"Find me highly-rated plumbers in Denver, CO"

The agent will call angi_search_providers with service_type: "plumber" and location: "Denver, CO".

Get provider details

"Show me reviews for this plumber: /pro/acme-plumbing-123/"

The agent will call angi_provider_profile with the URL.

Request a quote

"Request quotes for bathroom tile installation in zip code 80203"

The agent will call angi_request_quote with your project details.

Book an appointment

"Book a house cleaning appointment for next Saturday morning"

The agent will call angi_book_appointment with your preferred date and time.

Check your projects

"What home projects do I have going on right now?"

The agent will call angi_get_projects.

Authentication Flow

Angi uses browser-based authentication:

• Call angi_login — the MCP server opens Angi's login page in a headless browser
• Complete login manually (Angi may require email/password or Google sign-in)
• Call angi_status to confirm you're logged in
• Your session is saved to ~/.strider/angi/ and reused automatically in future sessions

Session Storage

~/.strider/angi/
├── cookies.json    # Browser cookies (auto-managed)
└── session.json    # Session metadata

Technical Details

• Uses Patchright (Playwright with stealth patches) for browser automation
• Headless Chrome — no visible browser window
• Cookie persistence for seamless re-authentication
• Graceful cleanup on server shutdown

Limitations

• Requires manual login (Angi does not provide a public API)
• UI scraping may break if Angi updates their website layout
• Some features (booking, messaging) require an active Angi account

Troubleshooting

"Not logged in" errors Run angi_login, complete login in the browser, then call angi_status.

"Browser failed to launch" Ensure Playwright browsers are installed: npx patchright install chromium

Empty search results Try a more specific service type or verify the location format (city, state or zip code).

Development

# Install dependencies
npm install

# Build
npm run build

# Run directly
npm start