MOTHERDUCK BLOG

Vibe Coding has come for BI

The agent writes the SQL and the data viz now. Here's what's still your job — and how to implement data-layer changes that took a very hard natural language to SQL benchmark from 30% to 93% accuracy with Gemini Flash.

Internal vs. External Storage: What's the Limit of External Tables?

25 years of external tables, from Oracle 9i to DuckLake. When to use internal storage vs. external tables, why they keep getting re-added to every major data warehouse, and a benchmark showing the speed tax for cold data.

MotherDuck on Cloudflare Workers

Build a real-time voting app on Cloudflare Workers that queries MotherDuck through the Postgres endpoint, using Durable Objects for live state and the pg package for analytical workloads.

MotherDuck Skills: Teaching Your AI Agents to Do Analytics

MotherDuck Agent Skills is an open-source catalog that teaches AI coding agents how to work with MotherDuck: exploring schemas, writing DuckDB SQL, using the REST API, and building Dives. If MCP gives agents hands, skills give them a playbook.

DuckDB Ecosystem Newsletter : April 2026

DuckLake 1.0 ships, Lance adds vector search, Rust-native extension

Announcing DuckLake 1.0 on MotherDuck

MotherDuck now supports DuckLake 1.0, the open table lakehouse format designed for simplicity and low latency. Learn what's new in the 1.0 release, including data inlining, clustering, bucket partitioning, geometry and variant types, plus multi-engine support. Learn how DuckLake compares with Apache Iceberg and Delta Lake.

Water Town: The Agent Swarm Data Stack

In a fully agentic world, will we still need analytics at all? A particularly unhinged example might offer some clues.

Your AI dashboard looks cool. Nobody learns anything from it

Five fundamental data visualization principles that transform AI-generated dashboards from pretty charts into actionable insights. Learn chart selection, design, narrative, and interactivity.

Building a Text-to-SQL Agent with DuckDB, MotherDuck and LangChain

Learn how to build a text-to-SQL LLM and LangChain SQL agent with DuckDB and MotherDuck, including schema inspection, query checking, and self-correction.

Consulting the Oracle: Claude on the Future of Data

What does AI think its own impact will be on the data and analytics industry? I asked Claude to share its predictions on the Modern Data Stack, data jobs, and the future of analytics.

Embedded Dives: Interactive Data Apps for Customer-Facing Analytics

Real-time, interactive data apps – now embedded anywhere in your application. Created by your favorite agents and powered by dual execution.

Introducing HumanDB: The World's First Human-Powered Analytical Database

MotherDuck Now Speaks Postgres

MotherDuck's Postgres endpoint lets you query your data warehouse using any PostgreSQL-compatible client, driver, or BI tool — no new dependencies needed.

Specs Over Vibes: Consistent AI Results ft. Mark Freeman

Mark Freeman, Head of DevRel at Gable, shares his Spec-Driven Development workflow for producing high-quality, reproducible AI outcomes using Claude Code, ExcaliDraw diagrams, JSON schemas, and agent teams with tmux.

The Dive Gallery Is Live: Share Your Interactive Data Apps With the World

Share interactive, AI-built data apps with anyone through a unique link — queries run live on MotherDuck compute.

Don't Fear the Agents - AI on the Data Lakehouse

Learn how MotherDuck's Hosted DuckLake combines lakehouse scale (borrowing the best ideas from Apache Iceberg and Delta Lake) with low latency and agent sandbox isolation, making it the simplest data lakehouse for AI-powered agentic workflows.

Future Casting the Modern Data Stack

If the Modern Data Stack isn't yet dead, it's at least incredibly sleepy. AI is bringing the "long run" closer than ever — here's what might come next for ETL, BI, data warehouses, and the role of data engineers.

Claudeception: Inside the Mind of an Analytics Agent

More tool calls, more schema exploration, more verification — does it help, or hurt? We dug into the chain-of-thought traces behind one of the hardest text-to-SQL benchmarks to understand how analytics agents actually think.

DuckDB Ecosystem Newsletter – March 2026

SQL Transpilers, VS Code Extensions, Dives and more

DuckDB 1.5 Features I am Excited About

DuckDB 1.5 is both faster and easier to use! JSON analysis can be up to 100x faster thanks to the VARIANT type and JSON shredding. Real-world queries are faster, including both basic and complex queries. Writes to Azure are supported, and checkpoint concurrency is dramatically improved.

What is MCP? A Data Person's Guide to Agentic Analytics

MCP explained for data people: what it is, how it works, and how to connect your AI agent to your databases and tools

Dashboards as Code : CI/CD For MotherDuck Dives

Local development, version control, and automated CI/CD deployments for MotherDuck Dives — composable data visualizations that are just code.

Give Your Agents Write Access

The MotherDuck remote MCP server now supports write operations. Learn how zero-copy clones, snapshots, and hypertenancy make it safe to let agents build in your data warehouse.

Claude Code + Dives = Any data UI

Learn how to build custom interactive data visualizations using MotherDuck Dives and Claude Code. From setup to sharing, create refreshable React and SQL-powered visuals with natural language prompts.

How I dive - Claude.ai Edition

Explore first, find the story, iterate on the artifact, test in MotherDuck. The workflow I keep coming back to after weeks of building.

Git for Data Applied: Comparing Git-like Tools That Separate Metadata from Data

Comparing Git-like Tools That Separate Metadata from Data

Duck, Dive, and Answer

AI is eating the data stack. Introducing Dives, a new MotherDuck feature that lets AI agents build shareable, real-time data visualizations with composable SQL.

DuckDB Ecosystem Newsletter – February 2026

MSSQL Extension, Vortex File Format & 2TB Memory Benchmarks

Does "AI-Ready Data" simply mean "Good Data Modeling"?

I ran 500 text-to-SQL questions against three frontier LLMs with zero context beyond the schema. 95% accuracy. No semantic layer required.

Building an Obsidian RAG with DuckDB and MotherDuck

Build a local-first RAG for your Obsidian notes using DuckDB's vector search, then deploy it as a serverless web app with MotherDuck

More Control, Less Hassle: Self-Serve Recovery with Point-in-Time Restore

MotherDuck now supports point-in-time restores, making it easier than ever to roll back changes, undrop databases, and debug pipelines.

Microbatch: how to supercharge dbt-duckdb with the right incremental model

Microbatch support in dbt-duckdb lets you process large tables in time-based batches — making incremental models recoverable, backfillable, and parallelizable without rebuilding from scratch.

SQL Golf: Lessons from Quackmas 2025

From #N column references to boolean math, explore the extreme techniques used to solve the Quackmas 2025 SQL Golf challenge.

This Month in the DuckDB Ecosystem: January 2026

DuckDB news: Iceberg extension adds full DML (INSERT/UPDATE/DELETE). Process 1TB in 30 seconds. Query data via AI agents with MCP server. TypeScript macros for APIs.

We Built Our Employees a Wrapped—Using SQL and MotherDuck

We built a Spotify-style "Wrapped" for MotherDuck employees using our own platform. Discover the SQL queries, data tricks, and fun personas behind our internal leaderboard.

What If We Don't Need the Semantic Layer?

What if AI could discover your business logic by mining query history instead of requiring manual metric definitions? Explore how LLMs are replacing static semantic layers with systems that learn from actual usage.

A Quack-Packed Fall

MotherDuck spent fall on the conference circuit across Europe and the US. Here's what we presented and the pattern that emerged in nearly every conversation about analytics.

Stop Paying the Complexity Tax

My personal reflections from watching Jordan Tigani’s keynote "The Unbearable Bigness of Small Data" at Small Data SF. Most data warehouses are overbuilt for scale you'll never need. Learn how to stop paying the complexity tax.

Building the MotherDuck Remote MCP Server: A Journey Through Context Engineering and OAuth Proxies

How MotherDuck built a production-ready remote MCP server—from OAuth proxy challenges with Auth0 to tool design patterns that help AI agents query data warehouses effectively. Includes lessons from a hackathon where agents ran 4,000+

Building an answering machine

Discover how MotherDuck's MCP server enables true self-service analytics through AI agents. Query data in plain English with Claude, ChatGPT, or Gemini—no SQL required.

MotherDuck Integrates with PlanetScale Postgres

Run analytics 200x faster on PlanetScale Postgres with MotherDuck integration. Keep millisecond transactions while pushing analytical queries to serverless compute.

Calling All SQL Sleuths: The Christmas Heist Awaits

Christmas presents have vanished! Use your SQL skills to solve the mystery on DBQuacks and win prizes. 15 challenges. One leaderboard.

Streaming Pipelines with MotherDuck and Artie

CDC streaming takes flight with Artie’s new MotherDuck destination.

This Month in the DuckDB Ecosystem: December 2025

DuckDB news: v1.4 adds AES-256 encryption. DuckLake brings ACID-compliant lakehouse with time-travel queries. Gaggle extension queries Kaggle datasets directly via SQL.

Simplicity of a Database, but the Speed of a Cache: OLAP Caches for DuckDB

Speed up slow dashboards without adding new infrastructure. Learn how DuckDB's caching extensions can drop query times from minutes to seconds.

Branch, Test, Deploy: A Git-Inspired Approach for Data

This article explores how to bring Git style workflows like branching, testing, and deploying to your data stack. Learn how concepts like zero copy cloning and metadata pointers can finally give you isolated test environments.

Unstructured Document Analysis with Tensorlake and MotherDuck

Learn how to query unstructured documents with friendly SQL!

Small Data SF 2025: the Recap!

Dive into a recap of the world's hottest efficiency-themed data conference, Small Data SF!

DuckDB Ecosystem: November 2025

DuckDB news: QuackStore caching cuts query time from 49s to 3s. Infera runs ONNX ML models in SQL. 127 community extensions analyzed. DuckLake architecture explained.

4 Senior Data Engineers Answer 10 Top Reddit Questions

A great panel answering the most voted/commented data questions on Reddit

Faster Ducks

Benchmarks, efficiency, and how MotherDuck just got nearly 20% faster.

DuckDB 1.4.1 and DuckLake 0.3 Land in MotherDuck: New SQL Syntax, Iceberg Interoperability, and Performance Gains

MotherDuck now supports DuckDB 1.4.1 and DuckLake 0.3, with new SQL syntax, faster sorting, Iceberg interoperability, and more. Read on for the highlights from these major releases.