Use with AI
These docs are built to be consumed by AI assistants and LLM tools. Three access methods ship out of the box — an MCP server, llms.txt files, and per-page raw markdown.
MCP server
A Model Context Protocol server is exposed at /mcp (https://hydrasdk.com/mcp) over streamable HTTP, letting an AI assistant query the documentation directly instead of guessing. The server is named Hydra SDK and provides two tools:
| Tool | Purpose |
|---|---|
list-pages | List every documentation page with its title, path, and description — use it to explore or search when you don't know the exact page. |
get-page | Fetch the full markdown of a specific page by path (e.g. /getting-started/installation). |
A typical flow: call list-pages to discover the right path, then get-page to read it.
Connect
claude mcp add --transport http hydra-sdk https://hydrasdk.com/mcp
{
"mcpServers": {
"hydra-sdk": {
"url": "https://hydrasdk.com/mcp"
}
}
}
pnpm dev, port 3003), the MCP server is at http://localhost:3003/mcp, and the dev server prints an Install Nuxt MCP server deep link on startup — open it to register the server in your editor automatically.llms.txt
Following the llms.txt convention, two files give an LLM the whole documentation as context:
- /llms.txt — an index of every page with links and descriptions.
- /llms-full.txt — the entire documentation concatenated into a single file.
Point an LLM-aware tool at either URL, or paste the contents into a chat to give the model full context about Hydra SDK.
Raw markdown
Add .md under the /raw/ prefix to any page path to get its clean markdown source — handy for piping a single page into a model:
curl https://hydrasdk.com/raw/getting-started/installation.md
The raw endpoint is locale-aware — prefix the locale to fetch a translated page:
curl https://hydrasdk.com/raw/vi/getting-started/installation.md
curl https://hydrasdk.com/raw/ja/getting-started/installation.md