Connect Candydocs MCP with Claude, Cursor, and Cline

    Learn how to connect the Candydocs MCP server with AI assistants like Claude Desktop, Cursor, and Cline to automatically create, update, and manage documentation, roadmap items, and changelog updates using AI agents.

    Product Guide
    8 min read
    20 Mar 2026
    mcp

    Candydocs provides a Model Context Protocol (MCP) server that allows AI assistants and developer tools to interact directly with your documentation platform.

    By connecting Candydocs MCP with AI tools like Claude Desktop, Cursor, or Cline, you can allow AI agents to:

    • Create documentation articles

    • Update knowledge base content

    • Manage product roadmap items

    • Publish changelog updates

    • Handle feature requests and wishlist items

    • Manage custom pages

    This guide shows how to connect Candydocs MCP to popular AI agent tools.


    Candydocs MCP Server

    The Candydocs MCP server exposes documentation management tools through the MCP protocol.

    MCP Server Endpoint

    https://mcp.candydocs.com/mcp

    MCP Discovery Manifest

    https://mcp.candydocs.com/.well-known/mcp.json

    Tool Definitions

    https://mcp.candydocs.com/tools/list

    Prerequisites

    Before connecting Candydocs MCP, make sure you have:

    1. A Candydocs account

    2. An API key

    3. One of the supported AI tools:

    Supported AI tools:

    • Claude Desktop

    • Cursor AI

    • Cline

    • Custom MCP agents

    • OpenAI Agents SDK


    Step 1 — Get Your Candydocs API Key

    Login to your Candydocs dashboard.

    Navigate to:

    Settings → API Keys

    Generate a new API key.

    Example:

    pk_live_xxxxxxxxxxxxx

    This API key will be used for authentication when connecting to the MCP server.


    Step 2 — Understand MCP Authentication

    Candydocs MCP uses Bearer token authentication.

    All MCP requests must include:

    Authorization: Bearer YOUR_API_KEY

    The MCP server also uses session-based connections.

    When an MCP client connects:

    1. The client initializes a session

    2. The server returns a session ID

    3. Future requests include the session header

    mcp-session-id: SESSION_ID

    Sessions remain valid for 24 hours.


    Connect Candydocs MCP with Claude Desktop

    Claude Desktop supports MCP servers through its local configuration.

    Step 1 — Open Claude MCP Settings

    Locate the MCP configuration file.

    Mac:

    ~/Library/Application Support/Claude/mcp_servers.json

    Windows:

    %APPDATA%\Claude\mcp_servers.json

    Step 2 — Add Candydocs MCP Server

    Add the following configuration:

    {
      "mcpServers": {
        "candydocs": {
          "command": "npx",
          "args": [
            "mcp-remote",
            "https://mcp.candydocs.com/mcp"
          ],
          "env": {
            "AUTHORIZATION": "Bearer YOUR_API_KEY"
          }
        }
      }
    }

    Replace:

    YOUR_API_KEY

    with your Candydocs API key.


    Step 3 — Restart Claude Desktop

    Restart Claude.

    Claude will automatically discover the Candydocs tools.

    You can now prompt Claude like:

    Create a new article in Candydocs explaining our authentication API.

    Claude will call the MCP tools automatically.


    Connect Candydocs MCP with Cursor

    Cursor supports MCP through its AI tool integrations.

    Step 1 — Open Cursor Settings

    Open:

    Cursor → Settings → MCP Servers

    Step 2 — Add Candydocs MCP Server

    Add a new server configuration.

    Server URL:

    https://mcp.candydocs.com/mcp

    Authentication:

    Bearer YOUR_API_KEY

    Step 3 — Restart Cursor

    After adding the server, restart Cursor.

    Cursor will fetch the available tools automatically.

    You can now ask Cursor:

    Create a documentation article for our REST API using Candydocs.

    Connect Candydocs MCP with Cline

    Cline is a developer AI assistant that supports MCP tools.

    Step 1 — Open Cline MCP Configuration

    Locate the configuration file:

    .cline/mcp.json

    Step 2 — Add Candydocs MCP Server

    Add the server entry:

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

    Step 3 — Restart Cline

    Restart Cline.

    Cline will automatically discover all Candydocs tools.

    You can now run prompts like:

    Create a roadmap item for "AI-powered search feature".

    Cline will call the MCP server to create the roadmap entry.


    Example AI Workflows

    After connecting Candydocs MCP, AI tools can automate documentation tasks.

    Example prompts:

    Generate Documentation

    Write a tutorial about our API rate limiting and publish it in Candydocs.

    Update Existing Articles

    Update the authentication article and add OAuth examples.

    Manage Product Roadmap

    Create a roadmap item for "Dark Mode Support".

    Publish Product Updates

    Create a changelog entry for version 2.3 release.

    Candydocs MCP Tool Categories

    The MCP server exposes tools across several product areas.

    Category

    Description

    Articles

    Manage documentation and knowledge base

    Roadmap

    Track product development plans

    Updates

    Publish changelog updates

    Wishlist

    Manage feature requests

    Custom Pages

    Manage additional documentation pages


    Troubleshooting

    MCP tools not appearing

    Verify the server URL:

    https://mcp.candydocs.com/mcp

    Authentication errors

    Ensure the header is correct:

    Authorization: Bearer YOUR_API_KEY

    Session errors

    Make sure the MCP client performs:

    initialize

    before calling tools.


    Learn More

    Candydocs MCP endpoints:

    https://mcp.candydocs.com/.well-known/mcp.json
    https://mcp.candydocs.com/tools/list

    Summary

    By connecting Candydocs MCP with AI assistants like Claude, Cursor, and Cline, you enable powerful AI-driven workflows for managing documentation and product content.

    AI agents can:

    • write articles

    • update documentation

    • manage roadmap items

    • publish changelog updates

    • handle user feature requests

    All directly inside your Candydocs workspace.

    Need Help?

    Can't find what you're looking for?

    Comments (0)

    Sign in to comment

    You need to be logged in to post a comment on this article.

    No comments yet. Be the first to share your thoughts!