Developer resources

Clinic Directory for machines

Every page in this directory is available as Markdown, and the whole index is available as a sitemap and an llms.txt map. No account, no API key, no rate-limit negotiation.

Markdown content negotiation

Request any page with Accept: text/markdown and it is served as Markdown at the same URL. No redirect, no separate hostname, no .md suffix - the canonical URL stays canonical.

curl -H "Accept: text/markdown" https://clinicdirectory.net/austin/dentistry/

Coverage is automatic. Every page in the sitemap is available as Markdown, including new pages and new tools, because the renderer reads the page's own published HTML rather than a separate list. A page that is removed stops serving Markdown at the same moment it stops serving HTML.

This follows the four checks published at acceptmarkdown.com: Markdown is served for Accept: text/markdown, responses carry Vary: Accept, unsupported types get a 406, and q-values are honoured. Sending Accept: text/html;q=0.8, text/markdown;q=0.9 returns Markdown; reversing the weights returns HTML.

Directory map

/llms.txt - curated map of the directory following the llms.txt convention. Start here; it is small and hand-shaped.
/sitemap.xml - a <sitemapindex> pointing at chunked child sitemaps covering every clinic, city and specialty URL. Exhaustive and large.
/robots.txt - crawler policy, including Content-Signals declaring what this content may be used for.

URL patterns

Every directory URL is derivable. All paths carry a trailing slash.

/{city}/ - every clinic in a city, e.g. /austin/
/{city}/{specialty}/ - one specialty in one city, e.g. /austin/dentistry/
/{city}/{specialty}/{clinic-name}/ - a single clinic profile
/cities/, /states/, /specialties/ - A–Z indexes

City and specialty segments are lowercase, hyphenated slugs. Requesting a path that does not exist returns a real 404 with a Markdown or JSON body listing where to look instead, rather than a soft 200.

Error format

Machine-readable endpoints return a JSON envelope with a stable code to branch on and a hint describing the recovery path. Request any URL with Accept: application/json to get errors in this shape.

{
  "error": {
    "code": "not_found",
    "status": 404,
    "message": "No resource exists at /nope/",
    "hint": "Retrieve https://clinicdirectory.net/sitemap.xml for every valid URL.",
    "documentation": "https://clinicdirectory.net/developers/"
  }
}

Structured data

Clinic profiles emit MedicalClinic JSON-LD with address, telephone, services as an OfferCatalog, languages, and an AggregateRating where a plausible one exists. Listing pages emit BreadcrumbList and ItemList. Parsing the JSON-LD is usually cheaper than parsing the Markdown.

Attribution and limits

Clinic data is aggregated from each clinic's own website and public review platforms, and reflects what was published there at crawl time. It is not medical advice and is not a substitute for confirming details with the clinic. If you republish figures derived from this directory, describe them as describing this directory's coverage rather than the U.S. clinic population, and link the source page.

Questions, or need something this page does not cover? team@clinsight.ai