Getting started

Markdown

Using Externa with AI coding tools

Give coding agents current Externa docs instead of stale training data. Use Context7 for MCP-backed lookup, paste or fetch llms.txt / llms-full.txt, or ask on this site with Ask the docs.

API key

Context7 recommends a free API key from context7.com/dashboard for higher rate limits. Replace YOUR_API_KEY in the snippets below. Config below matches Context7’s MCP Clients docs.

Context7

Context7 is a service that serves up-to-date library documentation to AI coding assistants through an MCP server. Your agent calls Context7 tools instead of guessing from training data.

Externa is indexed under two library IDs:

Library IDSourceWhen to use
/qiick-io/externa-coreexterna-core package source and READMEImplementing or extending the Laravel app — routes, install, admin features, package internals
/llmstxt/externa_qiick_io_llms_txtThis docs site, indexed from https://docs.externa.qiick.io/llms.txtOperator and integration guides — Public CMS API, deployment, field types, headless starters

Prefer /qiick-io/externa-core for app/source questions. Prefer /llmstxt/externa_qiick_io_llms_txt for published docs pages and how-tos. Either works for many tasks; pin an ID when you already know which corpus you want.

Quick setup

Fastest path (Node.js 22.13+):

npx ctx7 setup

Use --cursor, --claude, or see Context7 install for other clients. Manual MCP configs follow.

Cursor

Go to Cursor Settings → Tools & MCP → New MCP Server, or edit ~/.cursor/mcp.json (or .cursor/mcp.json in a project).

Remote server (recommended):

{
  "mcpServers": {
    "context7": {
      "url": "https://mcp.context7.com/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Local server:

{
  "mcpServers": {
    "context7": {
      "command": "npx",
      "args": ["-y", "@upstash/context7-mcp", "--api-key", "YOUR_API_KEY"]
    }
  }
}

Claude Code

Local server:

claude mcp add --scope user context7 -- npx -y @upstash/context7-mcp --api-key YOUR_API_KEY

Remote server:

claude mcp add --scope user --header "Authorization: Bearer YOUR_API_KEY" --transport http context7 https://mcp.context7.com/mcp

Drop --scope user to install for the current project only. Or run npx ctx7 setup --claude.

VS Code

Add this to .vscode/mcp.json (see VS Code MCP docs).

Remote server:

{
  "servers": {
    "context7": {
      "type": "http",
      "url": "https://mcp.context7.com/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Local server:

{
  "servers": {
    "context7": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@upstash/context7-mcp", "--api-key", "YOUR_API_KEY"]
    }
  }
}

Context7 also ships a VS Code extension that registers the server and accepts an API key via context7.apiKey.

Sample prompts

Add use context7 so the agent fetches docs:

Create an Externa collection for blog posts and fetch it from Next.js. use context7

Pin a library ID to skip resolution:

Create an Externa collection for blog posts and fetch it from Next.js. use library /qiick-io/externa-core

Docs-site corpus:

How do I grant the public role Read on a collection for the Public CMS API? use library /llmstxt/externa_qiick_io_llms_txt

llms.txt and llms-full.txt

This site publishes LLM-oriented Markdown at build time (npm run build:llms):

FileURLRole
llms.txthttps://docs.externa.qiick.io/llms.txtIndex of every public docs page with title, short description, and link to a per-page .md mirror
llms-full.txthttps://docs.externa.qiick.io/llms-full.txtFull concatenated docs body — paste into an LLM or feed tools that ingest a single corpus

Individual page mirrors live at https://docs.externa.qiick.io/docs/<slug>.md (home: /index.md). Context7’s /llmstxt/externa_qiick_io_llms_txt library is built from the llms.txt index.

Tip

Prefer llms.txt when the tool can follow links. Use llms-full.txt when you need one pasteable dump or a tool that only reads a single file.

Ask the docs (on this site)

The header includes an Ask the docs control next to search (shown on browsers that support on-device generation). It answers from this documentation in the browser — no server-side chat backend. Use it for quick questions while reading; use Context7 or llms.txt when your coding agent needs Externa context in the IDE.