Agent-ready directory

Built for AI agents

LandscaperSearch works for people and for AI assistants. It hands agents clean, structured data and lets them call real tools here, instead of scraping and guessing from the page. Both capabilities are shown below, live, so you can verify them yourself.

Checking your browser…

Why it matters: when a directory is machine-readable, an AI assistant can answer “find me a landscaper in Barnet” with verified, structured data (services, coverage, ratings and hours) and cite this site, rather than inventing an answer. The same is true for a listed business: list or claim your business and it gets the same agent-ready treatment automatically.

1. Clean content for agents

Every page has a token-light Markdown twin. A normal browser gets the HTML page; an agent that asks for Accept: text/markdown gets the same page as clean Markdown at the same URL. There is also an llms.txt index that points agents at the whole directory.

2. Callable tools (WebMCP)

The site registers 6 tools an AI agent can call directly through the WebMCP standard (document.modelContext). Press Run on any tool to see exactly what an agent receives. The responses below are the real ones; live registration in the browser needs Chrome 149+ or Edge, but the outputs are the same either way.

search_landscapers location, service

Search the directory by location and service. Returns up to 10 matches with town, rating, listing URL and phone.

get_business_details business

The full listing for one business: contact, services, areas, opening hours, rating and links. Looked up by name or listing id.

list_locations (no input)

Every location covered, with listing counts and page URLs.

list_services (no input)

Every service in the directory, with counts and page URLs.

get_directory_stats (no input)

Headline totals: listings, verified partners, locations, services, and how to request a quote.

request_quote business, name, phone (+ email, postcode, message)

Submits a real quote request to a chosen business straight into the enquiry pipeline. An agent calls this only after confirming the details with you.

Open an example quote form →

Verify it independently

  • Google PageSpeed Insights. Run a report on this site: the “Agentic Browsing” section lists the registered WebMCP tools and validates their schemas. Open PageSpeed Insights →
  • Chrome's Model Context Tool Inspector. Google's official extension lists the tools and runs them live in Chrome 149+. WebMCP documentation →
  • The Markdown and llms.txt above. Open them in any browser, or from a terminal run curl -sH "Accept: text/markdown" https://www.landscapersearch.co.uk/business/cleo-m-gardening-barnet/ | head.