MCP Integration

Connect your AI assistant to Docjays documentation using the Model Context Protocol (MCP).

What is MCP?

The Model Context Protocol (MCP) is a standard for connecting AI assistants like Claude to external data sources. With MCP, your AI assistant can:

  • Read and search your documentation directly
  • Propose changes and improvements to documents
  • Answer questions based on your specific documentation
  • Access grounded knowledge and decision records

Supported AI Assistants

Docjays works with multiple AI assistants that support MCP:

Claude Desktop

Official Claude desktop application

macOSWindows

Cursor

AI-first code editor with built-in MCP support

macOSWindowsLinux

Windsurf

Collaborative IDE with MCP integration

macOSWindowsLinux

Claude Code CLI

Terminal-based AI assistant

macOSWindowsLinux

VS Code (Continue)

Continue extension for Visual Studio Code

macOSWindowsLinux

Setup Instructions

Follow these steps to connect Claude Desktop to your Docjays project:

Step 1

Generate an API Key

First, you need an API key to authenticate your AI assistant with your project:

  1. Go to your Project Settings → API Keys
  2. Click "Generate New API Key"
  3. Give it a name (e.g., "Claude Desktop")
  4. Copy the API key (you won't see it again!)

Important: Save Your API Key

The API key is only shown once. Store it securely (e.g., in a password manager).

Step 2

Locate Config File

Find your Claude Desktop configuration file:

macOS:

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

Windows:

%APPDATA%\Claude\claude_desktop_config.json

Linux:

~/.config/Claude/claude_desktop_config.json
Step 3

Update Configuration

Add your Docjays project to the MCP servers configuration:

json
{
  "mcpServers": {
    "docjays": {
      "command": "curl",
      "args": [
        "-X", "POST",
        "-H", "Authorization: Bearer YOUR_API_KEY_HERE",
        "-H", "Content-Type: application/json",
        "https://your-domain.com/api/mcp"
      ]
    }
  }
}

Replace the following values:

  • YOUR_API_KEY_HERE with your API key from Step 1
  • your-domain.com with your Docjays instance URL
Step 4

Restart Claude Desktop

Close and reopen Claude Desktop for the changes to take effect.

✅ Verification:

Once connected, you should see "docjays" listed in Claude's available tools. You can ask Claude to search your documentation or read specific documents.

Other IDE Setup

Configuration steps are similar for other AI assistants. Each has its own config location:

Cursor

Config Location: Settings → MCP Servers

Open Cursor settings, navigate to MCP Servers, click "Add Server", and paste your configuration.

Windsurf

Config Location: MCP → Servers → Add Server

Navigate to MCP settings, add a new server, and configure with your API key.

Claude Code CLI

Config Location: ~/.config/claude-code/mcp.json

Edit the MCP configuration file and add your Docjays server configuration.

VS Code (Continue)

Config Location: VS Code Settings → Extensions → Continue

Install the Continue extension, open its settings, and add your MCP server configuration.

Available MCP Tools

Once connected, your AI assistant can use these tools to interact with your documentation:

read_document

Read the contents of a specific document

Example:

"Can you read the API authentication document?"

search_documents

Search across all documents for specific keywords or topics

Example:

"Search for information about JWT tokens in my docs"

list_documents

List all available documents in the project

Example:

"Show me all the documents in this project"

propose_change

Propose an edit or improvement to a document

Example:

"Propose an update to improve the authentication guide"

query_grounded_knowledge

Query grounded knowledge modules for specific information

Example:

"What are the grounded decisions about our authentication approach?"

Example Prompts

Try these prompts with your AI assistant once MCP is configured:

Search

"Search my documentation for examples of error handling"

Read

"Read the architecture.md document and summarize the key decisions"

Analyze

"Review all API documentation and check for inconsistencies"

Propose

"Suggest improvements to the getting started guide based on best practices"

Troubleshooting

AI assistant can't see my documentation

  • • Verify your API key is correct in the config file
  • • Check that you restarted your AI assistant after updating the config
  • • Ensure your API key has the necessary permissions
  • • Check that the project URL is correct and accessible

Authentication errors

  • • API key format should be: dj_proj_<project-id>_<random-string>
  • • Make sure the API key hasn't been deleted or expired
  • • Regenerate a new API key if needed

Tools not appearing

  • • Check the config file syntax (must be valid JSON)
  • • Restart your AI assistant completely (not just the window)
  • • Check logs for error messages

Still having issues?

Check our FAQ page for more help, or reach out through your project settings.

Next Steps

Frequently Asked Questions

Find answers to common questions

Getting Started Guide

Learn the basics of Docjays

Was this page helpful?