Dedox

npx add-mcp 'https://api.dedox.ai/mcp'

MCP integration

Connect Claude or Cursor to your knowledge base.

The MCP server is experimental, so expect rough edges. Delete actions are optional when you create a key. Check what your assistant is about to do before you approve it.

What is MCP?

Model Context Protocol (MCP) is a standard protocol that enables AI assistants to interact with external systems and data sources.

MCP lets AI assistants like Claude Desktop, Cursor and ChatGPT access and manage your Dedox knowledge bases. Instead of copying content between apps, you can ask your assistant to:

• Create and update articles in your knowledge bases
• Search for specific content across all your knowledge bases
• Get analytics insights and performance metrics
• Manage knowledge bases and organise content
• Moderate comments and manage engagement

What people use it for

Ask your assistant when you need something. It only acts when you ask.

Draft from a source

Ask your assistant to draft an article from notes, a URL or pasted text. You still review and publish.

Bulk content creation

Create multiple articles at once from structured data (CSV, JSON, etc.). Useful when moving content from another tool.

Content migration

Migrate content from other platforms like Confluence, Notion, or documentation sites. Your AI assistant can help structure and organise the migrated content.

Analytics-driven content

Ask your assistant to look at views and searches, then suggest which articles to improve next.

Docs from other sources

Ask your assistant to draft docs from code comments, API notes or release notes when you need them.

Content updates

Ask your assistant to update an article when something changes. Nothing goes live until you publish.

Getting started

Connect your AI assistant to Dedox in just a few steps

1

Create an MCP key

Go to your Account Settings → Security and create an MCP key. Select the permissions you want to grant to your AI assistant.

Available Permissions:

• Articles: Read, Write, and optional Delete
• Knowledge Bases: Read, Write, and optional Delete
• Analytics: Read
• Comments: Read, Moderate
• Search: Read

2

Configure your AI assistant

Add the Dedox MCP server to your AI assistant configuration. The setup varies by assistant:

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "dedox": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://api.dedox.ai/mcp",
        "--header",
        "X-MCP-Key:YOUR_MCP_KEY_HERE"
      ]
    }
  }
}

Cursor

Configure via Cursor Settings UI or edit .cursor/mcp.json:

To access MCP settings: Press Cmd + Shift + J (Mac) or Ctrl + Shift + J (Windows), then navigate to Features → MCP. Your configured server will appear in the list.

Important: After clicking "Add to Cursor", you'll need to replace YOUR_MCP_KEY_HERE with your actual MCP key from your account settings.

{
  "mcpServers": {
    "dedox": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://api.dedox.ai/mcp",
        "--header",
        "X-MCP-Key:YOUR_MCP_KEY_HERE"
      ]
    }
  }
}

ChatGPT with MCP

Configure through ChatGPT's MCP settings interface, providing the Dedox server URL and your MCP key.

3

Start using MCP

Once configured, you can interact with your Dedox knowledge bases through natural language:

You:"Create a new article about API authentication in my documentation knowledge base"

AI:"I'll create that article for you. What should the title and content be?"

AI:"Here are your top 10 articles by views this month..."

What you can do

What the MCP server can do

Article management

• Create, update, and delete articles
• List articles with filters (status, category, tags)
• Get detailed article information
• Manage article metadata (SEO, categories, tags)

Knowledge base management

• Create and manage knowledge bases
• Update knowledge base settings and branding
• List all your knowledge bases
• Get knowledge base details and configuration

Analytics

• Get knowledge base analytics with date ranges
• View article-level performance metrics
• Analyse engagement and traffic patterns

Search and discovery

• Full-text search across articles
• Get all categories and tags used in knowledge bases
• Filter and organise content

Security and permissions

• Granular permission control per key
• Keys shown only once on creation
• Rate limiting based on subscription tier
• Audit logging

Frequently asked questions

What subscription tier do I need?

MCP access is available on Professional and Enterprise plans. If you're on a lower tier, you'll need to upgrade your plan to use MCP.

How secure are MCP keys?

MCP keys are hashed with SHA-256 before storage and are never stored in plaintext. Keys are shown only once on creation. You can create multiple keys with different permissions for different use cases, and delete keys at any time.

What are the rate limits?

Rate limits are based on your subscription tier:

• Professional: 1,000 requests per hour per key
• Enterprise: Unlimited

Can I use MCP with multiple AI assistants?

Yes. Create a separate key for each assistant. Each key can have different permissions, allowing you to control what each assistant can access.

Where can I find more technical documentation?

Check out our API documentation for detailed information about the MCP protocol implementation, all available tools, parameters, and examples.