AI Integration

MCP Server

Connect AI tools to this documentation portal via the built-in MCP server.

Experimental

About MCP

The Model Context Protocol (MCP) is an open protocol that creates standardized connections between AI applications and external services. This portal includes a built-in MCP server so AI agents can query Ontopix documentation during their workflows.

When connected, an AI tool can:

  • Search documentation proactively while generating responses
  • Retrieve full page content when it determines the documentation is relevant
  • Stay current with the latest published documentation

Server URL

The MCP server is available at:

https://docs.ontopix.dev/mcp

Available Tools

list-pages

Lists documentation pages with navigation structure. Supports filtering by path and badge maturity.

ParameterTypeDescription
pathstring (optional)Filter to a subtree (e.g., /schemas, /engineering)
recursiveboolean (optional)Return all pages in subtree (default: false)
includeBadgesstring (optional)Only show pages with these badges
excludeBadgesstring (optional)Hide pages with these badges

get-page

Retrieves the full Markdown content of a specific documentation page.

ParameterTypeDescription
pathstring (required)The page path (e.g., /docs/about)

get-collection-info

Get metadata about documentation collections, including source repository and issue URL.

ParameterTypeDescription
collectionstring (optional)Collection key for details (omit for overview of all)

Overview mode (no parameters): Returns all collections with page counts and badges.

Detail mode (with collection): Adds sections, source repository info, and issue URL for reporting documentation problems.

search-pages

Full-text search across documentation content — titles, headings, and body text.

ParameterTypeDescription
querystring (required)Search terms (e.g., "authentication patterns")
collectionstring (optional)Limit search to one collection
limitnumber (optional)Max results (default: 10, max: 50)

Badge Filtering

Pages and collections carry badges indicating maturity and status. Use badge filtering in list-pages to focus on the right content:

BadgeMeaningTypical use
productionStable, approved for useDefault for trusted guidance
betaFeature complete, still testingUse with awareness
experimentalProof of conceptMay change or be removed
draftWork in progressNot yet reviewed
deprecatedBeing phased outLook for replacement

Example: To list only production-ready engineering standards:

{ "path": "/engineering", "includeBadges": ["Production"] }

Setup

Claude Code

claude mcp add --transport http ontopix-docs https://docs.ontopix.dev/mcp

Cursor

Create or update .cursor/mcp.json in your project root:

.cursor/mcp.json
{
  "mcpServers": {
    "ontopix-docs": {
      "type": "http",
      "url": "https://docs.ontopix.dev/mcp"
    }
  }
}

Visual Studio Code

Create or update .vscode/mcp.json:

.vscode/mcp.json
{
  "servers": {
    "ontopix-docs": {
      "type": "http",
      "url": "https://docs.ontopix.dev/mcp"
    }
  }
}

Windsurf

Navigate to Settings > Windsurf Settings > Cascade > Manage MCPs > View raw config, then add:

.codeium/windsurf/mcp_config.json
{
  "mcpServers": {
    "ontopix-docs": {
      "type": "http",
      "url": "https://docs.ontopix.dev/mcp"
    }
  }
}

Customization

The MCP server uses @nuxtjs/mcp-toolkit. Custom tools can be added in server/mcp/tools/:

server/mcp/tools/search.ts
import { z } from 'zod'

export default defineMcpTool({
  description: 'Search documentation by keyword',
  inputSchema: {
    query: z.string().describe('The search query'),
  },
  handler: async ({ query }) => {
    const results = await searchDocs(query)
    return {
      content: [{ type: 'text', text: JSON.stringify(results) }],
    }
  },
})
See the MCP Toolkit documentation for tools, resources, prompts, and handlers.