MCP Tools Reference

Kentik’s AI Advisor MCP Server exposes three tools for interacting with AI Advisor:

Note: This reference and the MCP Payload Reference are specifically for developers building custom AI agents or integrations with the Kentik MCP Server. Standard Model Context Protocol (MCP) clients like Claude Desktop handle these tools automatically in the background.

Rate Limits

Because the MCP server uses the AI Advisor REST API, standard rate limits apply. The MCP server automatically handles periodic polling to respect these limits.

IMPORTANT: Use the ask_followup tool instead of creating new sessions via ask_question whenever possible to minimize API calls and maintain context.

ask_question

Submits a new question to AI Advisor and creates a new conversation session.

Parameters:

prompt (string, required): Your question about the network

Example:

{
  "prompt": "Show me all devices that are currently down"
}

Returns:

AI-generated answer in Markdown format
Session ID for follow-up questions
Real-time progress updates during processing

Typical Processing Time: 5-30 seconds

ask_followup

Asks a follow-up question in an existing conversation, maintaining context.

Parameters:

session_id (string, required): UUID of the existing chat session.
Note: This value is dynamic and must be retrieved from the previous response.
prompt (string, required): Your follow-up question.

Example:

{
  "session_id": "019ab80d-f7a1-739b-99e0-803f501fdea9",
  "prompt": "What about their interface utilization?"
}

Returns:

AI-generated answer with conversation context
Updated session information

Note: The session_id is provided in the response from ask_question. Save it to continue the conversation.

get_session_history

Retrieves the complete conversation history for a session.

Parameters:

session_id (string, required): UUID of the chat session

Example:

{
  "session_id": "019ab80d-f7a1-739b-99e0-803f501fdea9"
}

Returns: Formatted conversation history with all questions and answers.