# AI agents: MCP server, Markdown and discovery

> How AI agents can search and read the Emolysis website: a read-only MCP server, Markdown versions of the pages and discovery documents, with no sign-in.

AI agents can search and read this website through a read-only MCP server, Markdown versions of its pages and standard discovery documents, without signing in. Emolysis itself is in private development, and nothing is released yet.

## MCP server

The website runs a Model Context Protocol (MCP) server at `https://emolysis.com/mcp`. It serves public content only and changes nothing. It needs no sign-in, keeps no session and answers each request with one JSON response.

| Setting | Value |
| --- | --- |
| Endpoint | `https://emolysis.com/mcp` |
| Transport | Streamable HTTP, `POST` only |
| Protocol versions | 2025-11-25, 2025-06-18, 2025-03-26 |
| Authentication | None |
| Server card | [/.well-known/mcp/server-card.json](https://emolysis.com/.well-known/mcp/server-card.json) |

### Tools

| Tool | What it returns |
| --- | --- |
| `search_site` | The pages that best match a `query`, up to `limit` (1 to 10, default 5), with title, URL, description, section and the URL of the Markdown version. |
| `get_page` | A page's Markdown version, for a `path` such as `/docs/introduction` or a full URL on emolysis.com. |
| `list_pages` | Every page, or the pages of one `section` (product, shapes, platforms, integrations, docs, tutorials, glossary, news, faq, company, legal), with title, URL and description. |
| `about_emolysis` | Facts about Emolysis, Bandpey GmbH and its founder, the same facts as the site's structured data and llms.txt. |

Every tool only reads this website.

### Try it

An MCP client needs the endpoint URL and nothing else. To see a call without one:

```bash
curl https://emolysis.com/mcp \
  -H 'Content-Type: application/json' \
  -H 'Accept: application/json, text/event-stream' \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"search_site","arguments":{"query":"golden tests"}}}'
```

Pages of this website may call the server from a browser; a browser request from another website gets status 403. A request body may be up to 64 KiB.

## Markdown

Docs pages, FAQ answers, glossary terms, news posts, tutorials, the home page, the about page, the founder page and this page each have a Markdown version. There are two ways to get it:

- Add `.md` to the page's URL, as in [/docs/introduction.md](https://emolysis.com/docs/introduction.md). The home page's version is [/index.md](https://emolysis.com/index.md).
- Request the page's own URL with the header `Accept: text/markdown`. The answer is the same Markdown, with `Content-Type: text/markdown`, `Vary: Accept`, an `x-markdown-tokens` header (an estimate: the length in characters divided by four) and a `Link` header that names the HTML page as canonical.

HTML stays the default, and a page without a Markdown version always answers with HTML.

```bash
curl -H 'Accept: text/markdown' https://emolysis.com/docs/introduction
```

[/llms.txt](https://emolysis.com/llms.txt) lists the pages with a line on each, and [/llms-full.txt](https://emolysis.com/llms-full.txt) is a longer summary of Emolysis.

## Discovery

| File | What it is |
| --- | --- |
| [/.well-known/mcp/server-card.json](https://emolysis.com/.well-known/mcp/server-card.json) | The MCP server card: name, version, endpoint, transport and protocol versions. |
| [/.well-known/api-catalog](https://emolysis.com/.well-known/api-catalog) | An API catalog (RFC 9727). It lists the MCP server alone, with its server card and this page. |
| [/.well-known/agent-skills/index.json](https://emolysis.com/.well-known/agent-skills/index.json) | An agent skills index with one skill, `emolysis`: how to answer questions about Emolysis from this website. |
| [/auth.md](https://emolysis.com/auth.md) | Authentication for agents: there is none, and none is needed. |
| [/llms.txt](https://emolysis.com/llms.txt) | An index of the website for language models. |
| [/robots.txt](https://emolysis.com/robots.txt) | Crawler rules and content signals: search engines and AI systems may read, cite and learn from every public page. |
| [/sitemap.xml](https://emolysis.com/sitemap.xml) | Every public page, with the date it last changed. |

The home page's `Link` header names /llms.txt, the sitemap, the home page's Markdown version and the API catalog, so an agent finds them without reading HTML.

## Not offered

- No sign-in or accounts for agents, and no OAuth or OpenID metadata. Emolysis accounts are by invitation and not open yet.
- No public API for agents, no agent-to-agent card, no payments or checkout, and no action that changes anything.
- This server reads the website only. The Emolysis engine's own MCP server, which lets an editor or coding agent run Emolysis commands, is a different thing: see [The MCP server](https://emolysis.com/docs/mcp-server).

Questions or problems: hello@emolysis.com.

---

Page: https://emolysis.com/agents. Emolysis is a product of Bandpey GmbH, Munich, Germany. It is in private development, and nothing is released yet.