Kalpana MCP Documentation

Connect ChatGPT, Claude, Cursor, or any MCP-compatible agent to Kalpana to start searching templates, verifying assets, and running creative batches.

Prerequisites

Before connecting your agent:

You must have an active account on Kalpana.
You must belong to at least one workspace with templates:read permission.
To generate files, ensure your workspace has available rendering credits.

Verify Gateway Availability

Verify the hosted MCP gateway is reachable:

curl -i https://mcp.kalpana.one/health

Expected response:

{
  "status": "ok"
}

Connect Your Client

ChatGPT connects to Kalpana using the Client ID Metadata Document (CIMD) flow with zero manual client secret configuration:

Enable Developer Mode

In ChatGPT, navigate to Settings → Security and login → Developer mode and enable it.

Register MCP Endpoint

Open https://chatgpt.com/plugins, select the + (Add) button, and enter:

https://mcp.kalpana.one/mcp

Authorize with Kalpana

ChatGPT will automatically discover the OAuth metadata at /.well-known/oauth-authorization-server and prompt you to log into Kalpana. Approve the requested workspace scopes.

Start a Conversation

Open a new conversation and ask:

"List my accessible Kalpana workspaces."

Claude Desktop supports Streamable HTTP MCP connections directly in its configuration file.

Open Configuration

Open your Claude Desktop configuration file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json

Add Kalpana MCP Entry

Add kalpana under mcpServers:

{
  "mcpServers": {
    "kalpana": {
      "url": "https://mcp.kalpana.one/mcp"
    }
  }
}

Restart Claude Desktop

Restart Claude Desktop. When launching, Claude will initiate authentication and display the available Kalpana tools.

Cursor features built-in MCP server connectivity for project and global agent interactions.

Open Cursor Settings

Open Cursor and navigate to Settings → Features → MCP Servers.

Add New MCP Server

Click + Add New MCP Server and enter:

Name: Kalpana
Type: sse or http
URL: https://mcp.kalpana.one/mcp

Authenticate & Verify

Complete authorization when prompted. The status indicator will turn green once connected.

If you are building custom agents or integrating via Python/TypeScript SDKs:

curl -X POST https://mcp.kalpana.one/mcp \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer <your-oauth-token>" \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "tools/list",
    "params": {}
  }'

First Interaction Examples

Once connected, prompt your AI agent:

1. Check Workspaces

"Show me my accessible Kalpana workspaces and permissions." The agent will call list_workspaces and return your workspace names, roles, and allowed capabilities.

2. Search Templates

"Find a Kalpana template for a product sale or summer promotion." The agent will execute search_templates(query="summer promotion") and show previews with template dimensions.

3. Inspect Variables

"What fields can I customize on this template?" The agent will call get_template_inputs(templateId="...") and report all customizable text, image, and style variables.

Credit Spending Safeguard: The agent will never automatically render without your approval. It will first call create_batch to report the exact credit cost and prompt you for confirmation before calling run_batch.

Quickstart