Overview
Databricks Genie MCP Server is a Model Context Protocol (MCP) service that bridges large language models with the Databricks Genie API. It enables LLMs to interact with Genie spaces by asking natural language questions, starting new Genie conversations, and following up in ongoing sessions. The server can list available Genie spaces and fetch metadata such as titles and descriptions, empowering users to select the right space. It supports initiating conversations with natural language queries, and continuing them with follow-ups to refine results. For data retrieval, the server can retrieve SQL and result tables in a structured format, simplifying the incorporation of Genie data into LLM workflows. The project includes prerequisites and setup steps, including Python 3.7+, a Databricks workspace, a personal access token, and enabling the Genie API. It provides tooling to test via an inspector or Docker, and instructions for integrating with Claude Desktop. The MCP tools cover core operations: list spaces, get space info, start conversations, follow-ups, and retrieve results, enabling seamless, conversational access to Genie data from LLMs.
Features
get_genie_space_id()
List available Genie space IDs and titles.
get_space_info(space_id: str)
Retrieve the title and description of a Genie space.
ask_genie(space_id: str, question: str)
Start a new Genie conversation and get results.
follow_up(space_id: str, conversation_id: str, question: str)
Continue an existing Genie conversation.
get_results(space_id: str, ...)
Retrieve SQL query results and structured result tables.
Who Is This For?
- LLMs:Enable LLMs to query Genie spaces and retrieve SQL results via natural language.
- Data engineers:Integrate Genie data access into data pipelines and ML workflows.
- BI analysts:Query Genie data to power reports and dashboards with natural language prompts.
