Skip to main content

MCP Server

The MCP server (kadaikodi-mcp binary, from @kadaikodi/mcp) exposes Kadaikodi and Burdenoff Workspaces capabilities as Model Context Protocol tools for AI assistants like Claude Desktop, ChatGPT desktop, etc.

Install​

npm install -g @kadaikodi/mcp
# or
bun install -g @kadaikodi/mcp

Binary: kadaikodi-mcp

Run​

Stdio transport (typical for Claude Desktop):

kadaikodi-mcp --stdio

HTTP transport (for remote clients):

kadaikodi-mcp --http --port 3030

What it exposes​

The server publishes tools in two groups:

  • Kadaikodi tools — providers, workers, tasks, offerings, orders, reviews, investments, analytics
  • Burdenoff Workspaces platform tools — identity, workspaces, groups, files, tags, conversations, notifications, scheduling, integrations, export, billing

Tool names are prefixed with kadaikodi_*, so tools from different products can be mounted side by side without colliding.

Every tool call is made with your token and is permission-checked exactly like the equivalent CLI or API call.

Browse the live tool list​

kadaikodi-mcp --list-tools

Claude Desktop integration​

Add to your Claude Desktop config:

{
  "mcpServers": {
    "kadaikodi": {
      "command": "kadaikodi-mcp",
      "args": ["--stdio"],
      "env": {
        "KADAIKODI_WORKSPACE_ENDPOINT": "https://graphqlworkspaces.burdenoff.com/workspaces/graphql",
        "KADAIKODI_GLOBAL_ENDPOINT": "https://graphql.burdenoff.com/global/graphql",
        "LOG_LEVEL": "error"
      }
    }
  }
}

Environment variables​

Defaults point at PROD. Override per environment:

VariableDefaultDescription
KADAIKODI_WORKSPACE_ENDPOINThttps://graphqlworkspaces.burdenoff.com/workspaces/graphqlWorkspace gateway
KADAIKODI_GLOBAL_ENDPOINThttps://graphql.burdenoff.com/global/graphqlGlobal gateway
KADAIKODI_API_KEY—API key auth (skips device code)
KADAIKODI_ACCESS_TOKEN—Pre-set access token
KADAIKODI_WORKSPACE_ID—Pre-select workspace
KADAIKODI_ORGANIZATION_ID—Pre-select organization
LOG_LEVELinfoLogging level

Alternatively, use BURDENOFF_ENV=local|alpha|prod. See Environment Selection.

Development​

cd ~/products/kadaikodi/kadaikodi-mcp
bun install
bun run dev           # Dev server (stdio, hot reload)
bun run server:http   # HTTP transport mode
bun run test          # Vitest wiring tests
bun run sanity        # typecheck + lint + format + build
bun run build         # Production build

Run as a binary after building:

node dist/index.js               # stdio transport
node dist/index.js --http --port 3030

Repository: github.com/algoshred/kadaikodi-mcp

Next steps​