---
title: "Agent access: MCP, markdown and structured data · Norilla Coast"
description: "How AI agents and developers can read and query the Norilla Coast guide: MCP server, /ask endpoint, markdown twins, llms.txt, Schema.org endpoints, EntityMap, OKF bundle and discovery files."
url: https://norillacoast.com/agents/
updated: 2026-11-09
publisher: Norilla Coast
---
For AI agents and developers

# Built to be read by machines too.

Everything on this site is available in formats agents can use directly. Here is every endpoint, grouped by what it does.

## Discovery and trust

| Resource | URL | Notes |
| --- | --- | --- |
| Sitemap | [`/sitemap-index.xml`](https://norillacoast.com/sitemap-index.xml) | Every page, with last-modified dates |
| robots.txt | [`/robots.txt`](https://norillacoast.com/robots.txt) | Content-Signal policy, sitemap, schemamap, EntityMap and Agentmap directives |
| llms.txt | [`/llms.txt`](https://norillacoast.com/llms.txt) | Curated index linking to markdown versions of every page |
| ARD catalog | [`/.well-known/ard.json`](https://norillacoast.com/.well-known/ard.json) | Agentic Resource Discovery v0.91 entries. Also at /.well-known/ai-catalog.json |
| OKF bundle | [`/okf/index.md`](https://norillacoast.com/okf/index.md) | Open Knowledge Format v0.2. Download: /okf.tar.gz |
| Content-Signal | [`/robots.txt`](https://norillacoast.com/robots.txt) | search=yes, ai-input=yes, ai-train=yes, also sent as a response header |
| IndexNow | [`/sitemap-index.xml`](https://norillacoast.com/sitemap-index.xml) | Changed URLs are pushed to IndexNow on every production deploy |
| RSS | [`/rss.xml`](https://norillacoast.com/rss.xml) | New and updated guides |
| DNS-AID | [`/agents/#dns`](https://norillacoast.com/agents/#dns) | SVCB records under \_agents.norillacoast.com advertise the MCP server |

## Content

| Resource | URL | Notes |
| --- | --- | --- |
| HTML pages | [`/`](https://norillacoast.com/) | Server-rendered, no JavaScript needed to read |
| Markdown twins | [`/villages/thirroul.md`](https://norillacoast.com/villages/thirroul.md) | Append .md to any page path |
| Accept: text/markdown | [`/villages/thirroul/`](https://norillacoast.com/villages/thirroul/) | Request any page with this header to get markdown |
| llms-full.txt | [`/llms-full.txt`](https://norillacoast.com/llms-full.txt) | The whole guide in one markdown file |
| Agent skill | [`/.well-known/agent-skills/index.json`](https://norillacoast.com/.well-known/agent-skills/index.json) | Agent Skills Discovery v0.2.0 index with SHA-256 digests |
| Guides feed | [`/guides/`](https://norillacoast.com/guides/) | Dated articles, published to RSS |

## Meaning

| Resource | URL | Notes |
| --- | --- | --- |
| Schema.org graph | [`/`](https://norillacoast.com/) | JSON-LD @graph on every page with stable @ids |
| Schema endpoints | [`/schema/villages/thirroul.jsonld`](https://norillacoast.com/schema/villages/thirroul.jsonld) | Each page’s graph as its own file |
| Schema feed | [`/schema-feed.jsonl`](https://norillacoast.com/schema-feed.jsonl) | Every entity on the site, one JSON-LD object per line |
| schemamap.xml | [`/schemamap.xml`](https://norillacoast.com/schemamap.xml) | Schema Feeds v0.1 map of the feed and endpoints |
| EntityMap | [`/entitymap.json`](https://norillacoast.com/entitymap.json) | EntityMap v1.0: entities, relations, evidence. Readable at /entitymap.html |

## Actions

| Resource | URL | Notes |
| --- | --- | --- |
| MCP server | [`/mcp`](https://norillacoast.com/mcp) | Streamable HTTP, no auth |
| MCP server card | [`/.well-known/mcp/server-card.json`](https://norillacoast.com/.well-known/mcp/server-card.json) | Also /.well-known/mcp.json and /.well-known/mcp/server-cards.json |
| MCP Registry | [`/server.json`](https://norillacoast.com/server.json) | Registry entry for com.norillacoast/guide |
| WebMCP tools | [`/agents/#webmcp`](https://norillacoast.com/agents/#webmcp) | Registered with document.modelContext in supporting browsers |
| /ask | [`/ask?query=ocean+pools`](https://norillacoast.com/ask?query=ocean+pools) | Natural-language search. OpenAPI: /openapi.json |

## How do you connect to the MCP server?

Add this to your MCP client configuration. The server is public, read-only and needs no key.

```
{
  "mcpServers": {
    "norilla-coast": {
      "type": "http",
      "url": "https://norillacoast.com/mcp"
    }
  }
}
```

| Tool | What it does |
| --- | --- |
| `search_guide` | Search the Norilla Coast guide to the northern Illawarra and return the most relevant passages, each with its source URL. |
| `list_villages` | List the nine villages from Stanwell Park to Bulli, north to south, with tagline, best-for, train station, ocean pool and coordinates. |
| `get_village` | Return the full guide for one village as markdown: things to do, where to swim and stay, getting there, FAQs. |
| `get_page` | Return a guide page as markdown. Pages: "/" (overview), "/weekend/" (3-day itinerary), "/getting-here/" (from Sydney), "/guides/when-to-visit/", "/guides/ocean-pools/". |

## How do you use /ask?

Send a question as `query`. The response is JSON with ranked `results`, each with a `url`, `name`, `score` and `description`. Add `mode=summarize` for a short extractive answer. Parameters follow NLWeb's `/ask`.

```
curl "https://norillacoast.com/ask?query=which+villages+have+an+ocean+pool"
```

To get any page as markdown:

```
curl -H "Accept: text/markdown" https://norillacoast.com/villages/thirroul/
```

## WebMCP tools

In browsers that support WebMCP, every page registers read-only tools with `document.modelContext.registerTool()`: `search_norilla_guide`, `get_village_guide`, `get_long_weekend_itinerary` and `get_travel_options_from_sydney`. The script loads only when the API exists, so other browsers download nothing.

## DNS-AID records

The MCP server is advertised in DNS under the `_agents` label, following the DNS-AID draft:

```
_index._agents.norillacoast.com.      3600 IN TXT  "agents=guide:mcp"
guide._mcp._agents.norillacoast.com. 3600 IN SVCB 1 norillacoast.com. alpn="mcp" port=443
guide._mcp._agents.norillacoast.com. 3600 IN TXT  "endpoint=https://norillacoast.com/mcp" "cap=https://norillacoast.com/.well-known/mcp/server-card.json"
```

## Agent access FAQs

### Can AI models use content from this site?

Yes. The Content-Signal policy allows search indexing, AI answers and grounding, and AI training. Please cite the page URL when you quote the guide.

### Does the MCP server need an API key?

No. It is public, read-only and stateless. Point any MCP client that supports Streamable HTTP at https://norillacoast.com/mcp.

### How current is the information?

The guide was last reviewed on 2026-11-09. Every page shows its last-updated date, and the same date appears as dateModified in its structured data.