Skip to main content
Connect AI clients to your Kioku knowledge base via the Model Context Protocol.

​
What is MCP?

The Model Context Protocol (MCP) is an open standard that lets AI applications call external tools. Kioku runs an MCP server with streamable-http transport at /mcp.

​
Configuration

​
Claude Desktop / Claude Code

Add to your MCP client config: 
{
    "mcpServers": {
        "Kioku": {
            "url": "https://api.kioku.chat/mcp",
            "headers": {
                "Authorization": "Bearer YOUR_JWT_TOKEN"
            }
        }
    }
}

​
Via CLI

kioku mcp
Prints ready-to-paste MCP server config JSON with your current auth token.

​
Local Development

{
    "mcpServers": {
        "Kioku": {
            "url": "http://localhost:9100/mcp",
            "headers": {
                "Authorization": "Bearer YOUR_JWT_TOKEN"
            }
        }
    }
}

​
How It Works

  1. AI client (Claude, Cursor) connects to Hivemind’s /mcp endpoint
  2. Authentication via Authorization: Bearer <token> header
  3. Company and user context extracted from MCP session _meta
  4. AI client can call any of the registered MCP tools
  5. Results are returned as structured data the AI can reason about
The MCP session is scoped to your company. All tool calls operate within your company’s knowledge base.