MCP Server

Model Context Protocol (MCP) is a universal translator that helps AI assistants to connect easily and use business applications​. To know more Click here.

Interact with DI applications as MCP tools for External AI Assistants through MCP Client.

Register Gathr application as a tool in MCP Server #

In Gathr, register application as a tool in MCP Server from Data Warehouse Intelligence apps listing page.

• Click ellipses of the listed application. Select MCP Server option.

• MCP Server window opens.

Provide the below details.

Register Gathr Data Warehouse Intelligence Application in MCP Server #

Prerequisites:

Interact with the Data Warehouse Intelligence application as a tool through your MCP Client.

Navigate to User Profile > Access Keys > Data Warehouse Intelligence API Keys to generate a token.

Provide HTTP request as a token for YOUR_API_KEY in MCP server.

For details see, Gathr DI MCP Integration

Connect to your MCP server - choose the MCP client to be used.

For details, see MCP Client

Enable checkbox to make this application available as a tool in your MCP client. #

Description #

Natural language explanation of what the tool does and how to use it. Helps the AI to understand when and how to call this tool.

RESET #

Option to reset the description as default.

GENERATE DESCRIPTION #

Option to generate an AI based description.

Connect to MCP Server #

MCP Client #

Choose the MCP client to connect with the MCP server.

Gathr DI MCP Integration #

Use the following settings to link your MCP client with available tools via the MCP server.

1. Replace YOUR_API_KEY with the API key you generated for Data Warehouse Intelligence in Gathr.

2. Optional: Use X-App-Names in header to restrict discovery to specified apps (comma-separated). Example: “X-App-Names”: “SalesApp,AnalyticsApp,MarketingApp”

3. Add the configuration to your MCP client (for example, Claude, OpenAI).

4. Discover the added tools by asking the assistant: “Use the discover_apps tool to show my available applications”.

AI Assistants #

Cursor #

Add the below configurations to your client’s MCP server settings.

• Open Cursor settings.
• Navigate to the “MCP Tools” tab and click New MCP Server.
• Copy/paste the following JSON configuration from below, then hit CMD+S or CTRL+S to save.
• To use Zapier MCP within Cursor, set the chat to Agent mode.

{
    "mcpServers": {
      "gathr-di": {
        "command": "npx",
        "args": [
          "mcp-remote",
          "https://app.gathr.co.in:8001",
          "--header",
          "Authorization: Bearer YOUR_API_KEY",
          "--header",
          "X-Project-Name: Default"
        ]
      }
    }
}

Claude Desktop #

Add the below configurations to your client’s MCP server settings.

• Open Claude Desktop
• Click on settings/preferences (gear icon)
• Look for “MCP Servers” or “Configuration” section
• Some versions have “Edit Config” button

If the config doesn’t exist then follow the below steps:

• Create new file named claude_desktop_config.json in the Claude folder

• Add your MCP server configuration

• Restart Claude Desktop

{
  "mcpServers": {
    "gathr-di": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://app.gathr.co.in:8001",
        "--header",
        "Authorization: Bearer YOUR_API_KEY",
        "--header",
        "X-Project-Name: Default"
      ]
    }
  }
}

API’s #

OpenAI Responses API #

Add the below configurations to your client’s MCP server settings.

• Get an OpenAI API Key

• Ensure that you have a valid OpenAI API key.

• To view or generate a new OpenAI API Key, go to the OpenAI platform API Keys page.

• Export your API key as an environment variable:

export OPENAI_API_KEY="sk-..."

import openai
 
client = openai.Client(api_key="your-key")
response = client.responses.create(
    model="gpt-4",
    input="Hello world",
    tools=[{
        "type": "mcp",
        "server_url": "http://localhost:8001/mcp",
        "headers": {
            "Authorization": "Bearer your-token",
            "X-Project-Name": "project_name"
        }
    }]
)
print(response)

curl -X POST https://api.openai.com/v1/responses \
  -H "Authorization: Bearer your-openai-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4",
    "input": "Hello world",
    "tools": [{
      "type": "mcp",
      "server_url": "http://mcp.gathr.ai:8001/mcp",
      "headers": {
        "Authorization": "Bearer your-token",
        "X-Project-Name": "project_name"
      }
    }]
  }'

The server_url may change as per gathr deployment unlimited version.

#Initialize OpenAI client
client = openai.Client(api_key=OPENAI_API_KEY)
response = client.responses.create(
	model=model,
	input=prompt,
	tools=[{
		"type": "mcp",
		"server_label": "gathr-di--server",
		"server_url": https://app.gathr.co.in:8001,
		"require_approval": "never",
		"headers": {
			"Authorization": f"Bearer YOUR_API_KEY}",
			"X-Project-Name": VamsiEmr
		}
	}]
)

Integrated Development Environment #

Windsurf #

Add the below configurations to your client’s MCP server settings.

Configuring Zapier MCP in Windsurf

• Open Windsurf settings (⌘+,)
• Navigate to the “Cascade” tab and click “Add Server”
• Select “Add custom server +”
• Add the following JSON configuration, replacing the placeholder URL with your MCP URL from below

{
  "mcpServers": {
    "gathr-di-server": {
      "serverUrl": "https://app.gathr.co.in:8001",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY",
        "X-Project-Name": "VamsiEmr"
      }
    }
  }
}

Other #

Apify MCP Client, VS Code MCP extension, custom implementations, and other MCP-compatible tools.

Provide the below details:

MCP Server URL #

Url of the MCP Server for updates. Use SSEClientTransport if the URL ends with /sse;otherwise, use HttpStreamableClientTransport.

MCP Transport Type Specification #

This setting helps you to override the MCP transport layer if required. Use SSEClientTransport for server Sent Events or HttpStreamableClientTransport for Streamable HTTP.

HTTP Headers #

HTTP Headers to be sent with the request to the MCP server. If you are using Apify’s MCP Server, headers are not required.

• Provide Server URL

https://app.gathr.co.in:8001

• Transport

http or streamable-http

• Authentication

Header-based

{
  "server_url": "https://app.gathr.co.in:8001",
  "transport": "http",
  "headers": {
    "Authorization": "Bearer YOUR_API_KEY",
    "X-Project-Name": "Default"
  }
}

Functions #

Supported functions on tools for an AI assistant.

Query Functions #

Get a response to a question in HTML, Markdown (MD), or Raw format.

Suggest Graph #

Get suggestions for graph details based on a query and data file path.

Get Data #

Retrieve data from the specified file path.

Discover Apps #

Get a list of MCP-enabled applications associated with the user’s API key.