Chat with a Duck

A while ago I published sql-workbench.com and the accompanying blog post called "Using DuckDB-WASM for in-browser Data Engineering". The SQL Workbench enables its users to analyze local or remote data directly in the browser.

This lowers the bar regarding the infrastructure needed to get started with Data Analysis and Data Engineering. No databases must be installed on servers or developer machines, no data is sent to the internet and SaaS or Cloud providers. The sole interface to the data you want to analyze is SQL.

But what if a user is new in this space, and has no SQL skills yet? Or, if he/she has SQL skills, but wants to query the data without having to understand the data model first?

LLMs (Large Language Models) can be used to generate code from natural language questions. Popular examples include GitHub Copilot, OpenAI's GPT-4 or Meta's Code Llama.

DuckDB-NSQL is a Text-to-SQL model created by NumberStation for MotherDuck. It's hosted on HuggingFace and on Ollama, and can be use with different LLM runtimes. There are also some nice blog posts that are worth a read:

Leave a Comment