Verdocs - AI & MCP Integration

Use Verdocs documentation with AI assistants and MCP-compatible tools like Claude Desktop, Cursor, and Windsurf.

LLM-Ready Documentation

The Verdocs Developer Center provides machine-readable endpoints that make our documentation accessible to AI assistants, coding tools, and automation workflows.

Documentation Index

https://developers.verdocs.com/llms.txt

Returns a lightweight index of every documentation page with its title, URL, and description. Use this as a starting point to discover what content is available.

Full Documentation Export

https://developers.verdocs.com/llms-full.txt

Returns the complete processed content of every documentation page in a single plaintext response. This is useful for ingesting the entire Verdocs knowledge base into an AI tool or RAG pipeline.

Per-Page Markdown

Any documentation page can be retrieved as processed Markdown by appending .mdx to the URL. For example:

https://developers.verdocs.com/docs/guides/tutorials/tagging-dynamic-documents.mdx

This is useful when you want to feed a specific page's content to an AI assistant without the surrounding site layout.

Verdocs documentation MCP (HTTP)

The site exposes a streamable HTTP MCP endpoint for documentation search and page fetch (backed by the same Next.js search index and markdown export as the rest of the site):

https://developers.verdocs.com/mcp

Tools

Tool	Purpose
`search_docs`	Full-text search over documentation titles and descriptions (same behavior as the site search API).
`get_doc_page`	Load processed markdown for a page. Pass a path such as `learn/mcp` or `docs/learn/mcp.mdx` (under the `/docs` site section); the server normalizes it to `/docs/...mdx`.

Clients must support MCP over HTTP (not only stdio). Configure the MCP base URL above in your client if it supports remote HTTP servers.

After wrangler dev or deploy, you can sanity-check the transport with MCP Inspector pointed at your /mcp URL, or with curl (response shape depends on negotiated protocol; you should see a JSON-RPC result or an SSE stream):

curl -sS -X POST "https://developers.verdocs.com/mcp" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -H "Mcp-Protocol-Version: 2024-11-05" \
  -d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2024-11-05","capabilities":{},"clientInfo":{"name":"curl","version":"1"}}}'

The Worker emits structured JSON log lines for tool invocations (event: mcp_tool) for observability. For abuse protection at scale, prefer Cloudflare WAF / rate limiting in front of the Worker; the Worker rejects bodies larger than 2 MiB on /mcp.

Using with MCP Clients (fetch-based)

The Model Context Protocol (MCP) also allows AI assistants to access static URLs. You can configure MCP-compatible tools to read Verdocs documentation directly without the HTTP MCP server.

Claude Desktop

Add the following to your Claude Desktop MCP configuration file (claude_desktop_config.json):

{
  "mcpServers": {
    "verdocs-docs": {
      "command": "npx",
      "args": [
        "-y",
        "@anthropic-ai/mcp-fetch@latest",
        "https://developers.verdocs.com/llms-full.txt"
      ]
    }
  }
}

You can then ask Claude questions about the Verdocs API, SDKs, and platform features using the fetched documentation as context.

Cursor / Windsurf

In Cursor or Windsurf, you can add https://developers.verdocs.com/llms-full.txt as a documentation source in your project's AI context settings. This allows the editor's AI assistant to reference Verdocs documentation when helping you write integration code.

Alternatively, add the URL to your project's .cursor/rules or equivalent configuration:

@docs https://developers.verdocs.com/llms-full.txt

If your editor version supports remote MCP over HTTP, add the documentation MCP URL (https://developers.verdocs.com/mcp) according to that product's current UI (options change frequently between releases).

Custom integrations

If you are building a custom MCP server that needs access to Verdocs documentation, you can fetch from any of the endpoints above. The per-page Markdown endpoint is particularly useful for targeted retrieval:

const page = 'guides/tutorials/tagging-dynamic-documents';
const response = await fetch(`https://developers.verdocs.com/docs/${page}.mdx`);
const markdown = await response.text();

Verdocs platform MCP (planned)

We plan a separate MCP surface that exposes Verdocs product API operations as tools (beyond reading this developer site). That will allow assistants to interact with the Verdocs platform on your behalf.

Planned capabilities include:

Template Management — List, search, and inspect templates in your organization
Envelope Operations — Create and send envelopes, check signing status, and retrieve completed documents
Recipient Management — Add recipients, generate signing links, and monitor signing progress
Organization Settings — Query API keys, webhook configurations, and entitlements

When available, you will be able to configure it in Claude Desktop or any MCP-compatible client. Follow What's New for announcements.

AI & MCP Integration