For agents & developers

Default regional home

The site root redirects here so bookmarks and agents can target a stable entry URL:

Canonical site URL

Use this origin for links and templates (no trailing slash):

Canonical listing URLs (recommended for agents)

The most stable URLs for no-JavaScript fetchers include full business listings in the initial HTML response:

Pattern: /{category_slug}/{city_slug} (e.g. canyon-lake-ca).

Search URLs

Primary pattern: text query in q. Always encode query values with encodeURIComponent before inserting into a URL. Location-less searches redirect to the Inland Empire catalog so raw HTML always includes listings.

Example (encoded):

Category & city filters

• Category only: https://nearme.tel/search?category={category_name} returns a 200 HTML response with listings prefetched for the Inland Empire (no redirect required). Canonical URL: https://nearme.tel/{category_slug}/inland-empire-ca.
• Typed search without location: https://nearme.tel/search?q={search_term} is server-prefetched with city=inland-empire-ca inline (200 response, no redirect).
• Near-me location: add &lat=…&lng=… (and optionally &radius=25) to both category and typed search URLs when you need geo-ranked results on /searchitself. Without coordinates, the server cannot infer the user's position. Mirror the same params on GET /api/search.
• API mapping: /search?category=… → GET /api/search?category=… (category browse, nearest first). /search?q=… → GET /api/search?q=… (typed search, best match). Do not send both category and q on the same API call.
• City filter: add &city={city_slug} (preferred, e.g. canyon-lake-ca) alongside q or category.
• API filter payload: server calls should carry city + state (or city_slug if the API parses it).
• Pagination: &page=2 (and higher) when results span multiple pages.

Markdown responses (no JavaScript)

Supported catalog URLs can return plain Markdown for agents that do not execute JavaScript. Use either query param or content negotiation:

• Append ?format=md to a supported URL.
• Send Accept: text/markdown on the request.

Examples:

Responses include a Link header pointing at the HTML canonical URL.

Business detail pages (fully server-rendered)

Listing detail URLs are the best agent entry point after you have a specific business. The initial HTML includes name, phone, address, hours, ratings, and — when cached — the "What reviewers say" Q&A block.

• Meta tags: each listing sets meta description, og:title, og:description, and Twitter card fields from review insights or business data.
• JSON-LD: HTML includes application/ld+json blocks for LocalBusiness, BreadcrumbList, FAQPage(from review Q&A), and related matches. Markdown fetchers that strip <script> tags will not see these — use raw HTML or ?format=md instead.
• Verify structured data: curl -sL 'https://nearme.tel/business/…' | rg application/ld+json

Model Context Protocol (MCP)

MCP servers typically expose tools (e.g. "search local businesses") that your host calls. You can implement a tool that builds the URLs above and returns fetched HTML or JSON from your own backend if you proxy our public pages — keep requests respectful, cache when possible, and stay within the beta limits above (500 searches/month).

Optional: set NEXT_PUBLIC_MCP_PACKAGE_URL in the deployment environment to show a link to your MCP server or npm package here.

OpenClaw & other agents

Configure your agent with permission to open or fetch nearme.tel catalog URLs. Prefer canonical listing paths or ?format=md when your fetcher does not run JavaScript. A minimal flow: build a search or listing link from the user's task, retrieve the page (HTML or Markdown), then summarize using the same review and match signals users see in the UI. Treat this path as beta and respect the 500/month search budget when automating.