Custom FAQs

Question 1

How is 'Custom' deployed and integrated with LLMs?

Accepted Answer

'Custom' is designed for seamless deployment to Databricks Apps using the Databricks CLI. Once deployed, it integrates with LLM clients like Claude Desktop and Cursor via the Model Context Protocol, allowing LLMs to discover and use its tools.

Question 2

What kind of LLM-callable tools does 'Custom' offer?

Accepted Answer

It provides two primary tools: 'get_stock_info' for retrieving real-time stock information from the Tokyo Stock Exchange, and 'run_query_on_databricks' to execute SQL queries directly against your Databricks SQL warehouses.

Question 3

What are the benefits of running this on Databricks Apps?

Accepted Answer

Running on Databricks Apps provides robust infrastructure, easy deployment, managed environments, and secure access to your Databricks SQL warehouses, making it ideal for scalable and reliable LLM tool integration.

Question 4

What is 'Custom' and what does it do?

Accepted Answer

'Custom' is a project that deploys a Model Context Protocol (MCP) server on Databricks Apps. It provides two LLM-callable tools for data interaction: fetching Tokyo Stock Exchange (TSE) stock information and executing SQL queries on Databricks SQL warehouses.

Question 5

Can I extend 'Custom' with my own custom tools?

Accepted Answer

Yes, 'Custom' is built to be extensible. You can easily add more custom tools by using the `@mcp.tool()` decorator within the application code, defining their functionality and descriptions for LLMs to utilize.

Custom

Custom

Key Features

Use Cases

Key Features

Use Cases