Using AI With Utila Docs

Utila's developer documentation supports several AI-assisted workflows that help you understand and call the API faster:

• llms.txt - machine-readable index of all published pages and API endpoints
• MCP server (https://docs.utila.io/mcp) - connect Cursor, VS Code, or Claude Code to search docs and inspect the Utila OpenAPI spec directly from your IDE
• AI Dropdown - copy any page as Markdown, or open Claude or ChatGPT with that page's context preloaded

These tools give AI assistants access to your published documentation and OpenAPI spec. They do not access your vault or the Utila API - for programmatic API access, authenticate with a service account.

LLM feed files

Fetch the complete documentation index at:

https://docs.utila.io/llms.txt

This is an auto-generated, always-current index of all Utila documentation pages and API endpoints - titles, descriptions, and links in a format optimised for LLMs. Use it to:

• Ingest into a custom GPT as a knowledge file
• Give an AI assistant a full map of Utila's documentation before asking questions
• Discover what pages exist before drilling into specific content

llms.txt contains links and summaries, not full page content. For full page content, use the MCP server (which can fetch individual pages) or append .md to any page URL.

Connecting to the Utila MCP server

The Utila MCP server at https://docs.utila.io/mcp exposes your IDE agent to Utila's documentation and OpenAPI specification. Once connected, your agent can:

Claude

To use the Utila MCP server with Claude:

1. Navigate to the Connectors page in the Claude settings.
2. Select Add custom connector.
3. Add the Utila MCP server:
   • Name: utila-docs
   • URL: https://docs.utila.io/mcp
4. Select Add.

Then, to access the MCP server in your chat:

1. Select the attachments button (the plus icon).
2. Select the Utila MCP server.
3. Ask Claude a question about Utila's documentation.

See the Model Context Protocol documentation for more details.

Claude Code

Run:

claude mcp add --transport http utila-docs https://docs.utila.io/mcp

To confirm:

claude mcp list

For a project-scoped config shared with your team:

claude mcp add --transport http --scope project utila-docs https://docs.utila.io/mcp

This writes to .mcp.json in your repo root. See the Claude Code documentation for more details.

Cursor

1. Use Command+Shift+P (Ctrl+Shift+P on Windows) to open the command palette.
2. Search for "Open MCP settings".
3. Select Add custom MCP - this opens the mcp.json file.
4. Add the following:

{
  "mcpServers": {
    "utila-docs": {
      "url": "https://docs.utila.io/mcp"
    }
  }
}

5. Reload Cursor (or restart MCP servers).
6. In Cursor's chat, ask "What tools do you have available?" to confirm the connection.

Config applies globally from ~/.cursor/mcp.json. For a project-scoped config shared with your team, use .cursor/mcp.json in your repo root instead. See Installing MCP servers in the Cursor documentation for more details.

VS Code

MCP in VS Code works through GitHub Copilot Chat. Create a .vscode/mcp.json file in your project and add:

{
  "servers": {
    "utila-docs": {
      "type": "http",
      "url": "https://docs.utila.io/mcp"
    }
  }
}

See the VS Code MCP documentation for more details.

AI Dropdown - chat with any page

Every page on docs.utila.io has a dropdown menu (next to the Copy Page button) with the following options:

Copy Page / View as Markdown exports the current page only - not the full site. For cross-doc questions, the MCP server's search tools are more effective.

When to use which approach