How to Use the New DuckDB UI – Fast, Free, and Local

2025/03/14

TL;DR: The DuckDB UI is a free local SQL notebook. Run duckdb -ui and you get a browser-based interface with a table explorer, autocomplete, and fast local file querying. You can also connect it to MotherDuck to access your cloud databases.

How to install and launch it

Install the DuckDB CLI (one-line curl, brew install duckdb on macOS, or other methods), then run duckdb -ui. DuckDB installs the UI extension automatically and starts a local web server. Click the link and you're in a SQL notebook in your browser.

The four main areas

The UI has a catalog browser on the left, a table explorer that shows column-level stats like min/max dates and value distributions, a SQL notebook with autocomplete and keyboard shortcuts (Cmd+Enter to run, Cmd+/ to toggle comments), and a results viewer with column sorting and filtering. You can also attach local or remote databases by path.

It handles big local files well

In the video, Mehdi loads a ~10GB Parquet file with 42.7 million rows and queries it in under a second. The column explorer updates instantly on the result set, so you can get a feel for the data without writing extra queries. He runs an aggregation across all 40M+ rows to find the most-shared domains on Hacker News, and that also comes back in under a second.

Connecting to MotherDuck

There's a "Sign on to MotherDuck" button in the UI. Click it, authenticate, and your cloud databases show up alongside your local ones. Mehdi queries the NYC service complaints dataset from the MotherDuck sample data to show how local and cloud workflows work together in the same interface.

Source code and feature requests

The DuckDB UI is open source at github.com/duckdb/duckdb-ui. If you want to try the cloud side, you can get started with MotherDuck for free.

0:00dug DB the inprocess allab database can

0:02run anywhere it supports multiple client

0:05languages and as today a significant

0:08portion of its usage come from Python

0:11and the dgdb CLI for some users the CLI

0:14can feel intim mediating especially if

0:16they prefer a user interface and let's

0:19be honest exploring data in the CLI can

0:22be

0:22cumbersome but dgdb in partnership with

0:25Mod duck just released a local dgdb UI

0:28so let's talk about how this can

0:30transform your local data exploration

0:32experience so to get started with DUI

0:34you can install the CLI there is a one

0:37command line curl that you can use if

0:39you're on Mac OS you can just do pro

0:41install dgdb and you have also other

0:44installation metod that I'll put in the

0:46description finally when you launch dgdb

0:49you can directly put

0:51dasi to launch DB UI what is happening

0:55behind the scene is that dug DB is

0:56installing and loading the dug DB UI

1:00extension dug DB has an extension

1:01mechanism for various things like

1:04reading over park or Iceberg and dug DB

1:07UI is one of it so a local web server

1:09has been launched and I can click on

1:11this link and I will get a local SQL

1:15notebook interface so we have four main

1:17areas in this UI the first one is here

1:20the catalog as you can see is pretty

1:23empty so you can run the example and

1:25here I'm going to create table based on

1:28a remote Park file you see now my table

1:31is created and if I click on it I'm

1:34going to have the second area of the UI

1:37which is the table Explorer this giv me

1:39direct information about each of the

1:42columns I can directly see what's the

1:44latest earliest date of this data set

1:47and various other Matrix on other type

1:50of column I can of course click and

1:52attach a local or remote database which

1:54just the path so here for example I'm

1:56attaching another database over uh I

2:00believe it's three and you see now I

2:01have the remote station database added

2:04with the station table and again I can

2:06click and see through the table Explorer

2:09so let's go a bit over the SQL notebook

2:11you can create multiple notebook name it

2:14and it will be stored locally you can

2:16expect all the fancy shortcut that you

2:18would have in any do book so if I do

2:21select mandin as stck I want to run this

2:24one I do command enter I can create

2:27other cell and of course you the arrow

2:30key to switch but cell I can toggle

2:33command by doing command slash and it's

2:36a toggle into a SQL command and of

2:39course we load stuff from the cloud but

2:42it works really fast with local file so

2:45here I'm going to create a table and you

2:48see I have auto complete on the command

2:50iur use as select command from a local

2:54parket file which is in the current

2:57directory so this parket file is

2:59actually almost 10 GB of data so now

3:02it's over there let's look at the table

3:04Explorer you see we have

3:0642.7 millions rows and so now if I'm

3:10exploring this data set a bit let's do

3:12100,000 lines for example you see that I

3:15have the nice UI where I can uh display

3:19which colon I want to display sort it

3:21out and I have also the colon Explorer

3:24which is based on the result of my query

3:26so you see here I can see I have a bit

3:29of sample from data up to 2008 and up to

3:33the latest of uh April 2024 and of

3:37course if you want to Crunch the data

3:39against search a data set it also super

3:42fast here I'm extracting the top uh

3:45domain shared on acar news from uh 2016

3:49to

3:512025 uh that's the data set and you see

3:54in less than a seconds I'm processing

3:56over more than 40 millions of Bros and

3:59getting the top domains which is get up

4:01then YouTube then medium.com as often

4:03articles that being shared on acur news

4:05finally you may have seen this button

4:07here right there which is sign on to

4:10Mother dock if you already have an

4:11account and you click there you can sign

4:13up and now you still have the local UI

4:16but as you can see you have access to

4:19all your Cloud database the one are are

4:21shared so for example you can start uh

4:23create the sample data which is database

4:26that any Modoc user get access to with a

4:29couple sample of data from acur new

4:31stock overflow or New York city so let's

4:34say if I'm quiry this database and the

4:36New York City service complaint to run

4:38and see what is the most complaint uh

4:41from New York City service and I see

4:43it's like heat hot water uh and the

4:46second is the noise all of that now

4:48leveraging the cloud compute and the

4:50cloud storage still using the same UI

4:52all right that's it's a smooth powerful

4:54and easy way to explore data locally

4:57with wbui if you have any feature

5:00request you can head over the gith up

5:02repository and in the meantime keep

5:04clacking and keep coding in the UI

5:09[Music]

FAQS

How do I install and launch the DuckDB UI?

Install the DuckDB CLI from duckdb.org/docs/installation, then run duckdb -ui in your terminal. DuckDB will install the UI extension automatically and open a SQL notebook in your browser.

Can the DuckDB UI handle large local files?

Yes. In this video, a ~10GB Parquet file with 42.7 million rows is loaded and queried in under a second. The UI runs locally using DuckDB's in-process engine, so performance depends on your machine rather than a network connection.

How do I connect the DuckDB UI to MotherDuck?

Click the "Sign on to MotherDuck" button inside the UI and authenticate. Your cloud databases will appear in the catalog alongside any local databases, and you can query both from the same notebook. Get started with MotherDuck for free.

Is the DuckDB UI free and open source?

Yes. The DuckDB UI is a free extension and the source code is available at github.com/duckdb/duckdb-ui. You can file feature requests or contribute there.

FAQs Lottie

Related Videos

"MCP: Understand It, Set It Up, Use It" video thumbnail

9:09

2026-02-13

MCP: Understand It, Set It Up, Use It

Learn what MCP (Model Context Protocol) is, how its three building blocks work, and how to set up remote and local MCP servers. Includes a real demo chaining MotherDuck and Notion MCP servers in a single prompt.

YouTube

MCP

AI, ML and LLMs

" Preparing Your Data Warehouse for AI: Let Your Agents Cook" video thumbnail

2026-01-27

Preparing Your Data Warehouse for AI: Let Your Agents Cook

Jacob and Jerel from MotherDuck showcase practical ways to optimize your data warehouse for AI-powered SQL generation. Through rigorous testing with the Bird benchmark, they demonstrate that text-to-SQL accuracy can jump from 30% to 74% by enriching your database with the right metadata.

AI, ML and LLMs

SQL

MotherDuck Features

Stream

Tutorial

"No More Writing SQL for Quick Analysis" video thumbnail

0:09:18

2026-01-21

No More Writing SQL for Quick Analysis

Learn how to use the MotherDuck MCP server with Claude to analyze data using natural language—no SQL required. This text-to-SQL tutorial shows how AI data analysis works with the Model Context Protocol (MCP), letting you query databases, Parquet files on S3, and even public APIs just by asking questions in plain English.

YouTube

Tutorial

AI, ML and LLMs