How does the pricing work?

Pricing has a free Lite tier. Paid plans (Business, Enterprise) have a base fee + usage. Usage is based on storage (per GB/month), compute time (per CU-hour), and features like Flights and AI functions.

Can I connect my existing tools?

Yes. MotherDuck integrates with dbt, Fivetran, Tableau, Looker, and connects via Python, Node.js, Go, JDBC, and a Postgres-compatible endpoint.

What is 'hypertenancy'?

It's MotherDuck's architecture that provisions a dedicated, isolated DuckDB instance (a 'Duckling') for every user, dashboard, or AI agent session. This eliminates resource contention and 'noisy neighbor' issues common in shared compute warehouses.

What compute instance types are available?

Instances range from Pulse (lightweight, bursty) to Giga (largest workloads). They are priced per compute unit-hour and can be sized independently for different tasks like backfills, dbt runs, or end-user queries.

Is my data secure? What compliance is offered?

MotherDuck is SOC 2 compliant and GDPR-ready. Enterprise plans include HIPAA BAA, AWS PrivateLink, and IP allowlisting for enhanced security.

How do AI functions work?

MotherDuck provides built-in SQL AI functions (like prompt() and embedding()). These are metered per 'AI Unit' and run directly within your queries, simplifying AI-powered analytics.

Directory/Analytics/Cloud Data Warehouse/MotherDuck

MotherDuck

MotherDuck is a serverless cloud data warehouse built on DuckDB, optimized for high-performance analytics. It features a hypertenancy architecture that provides every user or AI agent with an isolated compute instance, ensuring sub-second queries with no resource contention. The platform supports internal business intelligence, customer-facing analytics, and agent-driven tools, offering flexible compute sizes from lightweight Pulse to large-scale Giga instances. Key capabilities include AI-powered natural language querying via MCP, automated data pipelines with Flights, and interactive dashboards with Dives. It integrates with major data tools like dbt, Tableau, and Fivetran, and supports multiple programming languages including Python, Node.js, and Go. MotherDuck is designed for data engineers, analysts, and developers who need a scalable, cost-effective alternative to traditional warehouses.

Analytics

Visit website

Added on

Sep 22, 2026

Serverless DuckDBHypertenancyAI/MCP ServerAgent Pipelines (Flights)Embedded AnalyticsRead Scaling

Product information

Everything worth knowing about MotherDuck

The complete picture — from everyday features to the technical detail builders and IT folks dig for.

Core features

What it actually does

Serverless Data Warehouse

Built on DuckDB, providing fast, serverless analytics with no infrastructure to manage.

Agent-Native Tooling

Features like MCP Server for natural language queries, Guides for agent context, and Flights for Python pipelines.

Multiple Compute Instance Types

Choose from Pulse, Standard, Jumbo, Mega, and Giga instances sized for different workloads, from lightweight queries to massive jobs.

Shared & Differential Storage

Supports Apache Parquet, Iceberg, DuckLake, and cloud object storage with cost-effective, low-latency options.

Broad Language & Tool Integrations

Connect via Python, Node.js, Go, WebAssembly, PostgreSQL endpoint, dbt, Fivetran, Tableau, and more.

Role-Based Access Control

Manage user permissions with preset roles in Lite and both preset and custom roles in Business and Enterprise plans.

Technical capabilities

Hypertenancy Architecture

Every user, dashboard, and agent gets an isolated compute instance ('Duckling') for sub-second analytics with no resource contention.

Must watch videos

Scientific Computing on Multidimensional Arrays with DuckDB & XQL | Alex Merose

29:00

Tutorial29:00

Scientific Computing on Multidimensional Arrays with DuckDB & XQL | Alex Merose

Any N-dimensional array can be represented as a two-dimensional table—and that opens scientific datasets to SQL. Alex Merose demonstrates how XQL turns multidimensional scientific data into a tabular representation that can be queried with DuckDB. The talk introduces DuckDB-Zarr and Xarray-SQL and shows how SQL can improve scientific computing and data-science workflows. You’ll learn about: • Representing multidimensional arrays as relational tables • Querying scientific datasets with DuckDB • Working with DuckDB-Zarr and Xarray-SQL • Combining familiar SQL workflows with array-oriented computing • The ideas behind XQL Alex Merose is a Founding Member of Technical Staff at Open Athena. Recorded at the MotherDuck + DuckDB August Meetup on August 27, 2026 in San Francisco. The Q&A has been omitted from this standalone talk. XQL: https://xql.systems/ Event: https://luma.com/md-duckdb-aug MotherDuck: https://motherduck.com/ DuckDB: https://duckdb.org/ #DuckDB #MotherDuck #ScientificComputing #DataScience

7:51

Tutorial7:51

Why DuckDB 2.0 is faster

DuckDB 2.0 is faster but the biggest gains depend on how your data is shaped. See what changes in real workloads: 2–3× faster S3 reads, dramatically faster deep recursive queries, and a new VARIANT type that makes semi-structured data smaller and faster. In this video @mehdio benchmarks the DuckDB 2.0 alpha against 1.5.5 and explains the engineering behind async I/O, the rewritten recursive CTE engine, and VARIANT shredding. ☁️🦆 Start using DuckDB in the Cloud for FREE with MotherDuck : https://hubs.la/Q02QnFR40 📓 Resources Full blog and SQL examples: https://motherduck.com/blog/why-duckdb-20-is-faster/ Quack (DuckDB Client/Server protocol) video : https://www.youtube.com/watch?v=GZulGjfKPGM Install the DuckDB 2.0 alpha: https://duckdb.org/docs/installation/ ➡️ Follow Us LinkedIn: https://linkedin.com/company/motherduck X/Twitter : https://twitter.com/motherduck Blog: https://motherduck.com/blog/ 0:00 Three ways DuckDB 2.0 gets faster 0:38 Faster recursive CTEs 1:08 Walking deep parent-child data 1:32 DuckDB 1.5 vs 2.0 benchmark 2:47 VARIANT and shredding 3:47 VARIANT storage and query benchmarks 4:30 Async I/O explained 5:30 S3 benchmark results 6:08 More DuckDB 2.0 features 7:03 The new storage format 7:34 Try the DuckDB 2.0 alpha #DuckDB #DuckDB20 #DataEngineering #SQL #Database #Analytics #Parquet #S3 #JSON #OLAP #DataAnalytics

AWS Bought DuckLabs and users want their own agents

58:47

Tutorial58:47

AWS Bought DuckLabs and users want their own agents

AWS bought DuckLabs, the company employing the core team behind DuckDB—while DuckDB remains open source. What does the deal mean for AWS, MotherDuck, and the data stack building around DuckDB? In this episode, @mehdio and @dumkydewilde unpack the acquisition, then connect Iceberg and Lance, SQL-first context for agents, Spotify's LLM evaluation funnel, AI-generated data lineage, CodexBar, WalGit, query optimizers, and the fight over whether SaaS products should ship their own agents or let users bring one. All links and notes : https://motherduck.com/podcast 📓 Resources Why AWS Bought DuckLabs: https://www.linkedin.com/pulse/why-aws-bought-ducklabs-michael-driscoll-1gm7c/ Analyze with Iceberg, Serve with Lance: https://seaweedfs.com/blog/iceberg-lance-workflow/ Why your agent should manage its context through SQL: https://youtu.be/uMudSbZ79-A?t=825 Better Experiments with LLM Evals: https://engineering.atspotify.com/2026/5/better-experiments-with-llm-evals-a-funnel-not-a-fork CodexBar: https://github.com/steipete/CodexBar Architectural lessons for AI-powered data warehouses: https://x.com/josharosen/status/2095488762532745712?s=12 WalGit: https://github.com/tobi/walgit Optimizers: The Low-Key MVP: https://duckdb.org/2024/11/14/optimizers Give users an MCP, not another agent: https://x.com/SamBroner/status/2091876429621592426 ➡️ Follow Us LinkedIn: https://linkedin.com/company/motherduck X/Twitter: https://twitter.com/motherduck Blog: https://motherduck.com/blog/ 00:00 Welcome to EXPLAIN ANALYZE 01:15 Why AWS bought DuckLabs 11:14 Analyze with Iceberg, serve with Lance 16:30 Why your agent should manage context through SQL 22:09 Better experiments with LLM evals 27:26 CodexBar makes AI usage visible 33:06 Architectural lessons for AI-powered data warehouses 38:10 WalGit: one Git server binary, backed by object storage 42:55 Query optimizers: the low-key MVP 47:13 Do users want your agent, or their own? 58:36 Wrap-up #ExplainAnalyze #MotherDuck #DuckDB #AWS #DuckLabs #DataEngineering #AIAgents #Iceberg #Lance #LLMEvals #MCP #DataPodcast

Who uses it

Real use cases, no hype

Customer-Facing Analytics

Embed fast, interactive analytics directly into your SaaS application.

Internal BI & Data Warehousing

Modern, cost-effective warehouse for internal teams, replacing bloated setups.

AI/Agent-Driven Analysis

Use MCP server to let AI agents query data with natural language.

Data Pipeline Orchestration

Run scheduled Python jobs (Flights) for ELT, enrichment, and AI tasks.

Rapid Prototyping & Dev

Iterate locally with DuckDB, then deploy and scale instantly in the cloud.

Lightweight Data Lakehouse

Query Parquet/Iceberg files directly in object storage via DuckLake.

What's great

Blazing fast DuckDB analytics, sub-second queries.
Truly serverless; zero infrastructure to manage.
Generous free tier (Lite plan) for solo devs.
Deep integration with modern data stack (dbt, Fivetran, Tableau).

Technical strengths

Agent-first 'hypertenancy' prevents noisy-neighbor slowdowns.
Flexible compute tiers (Pulse to Giga) for variable workloads.
Built-in AI functions (prompt, embedding) and MCP server.

Where it falls short

Costs can escalate with compute-heavy usage.
Per-user limits on lower tiers (e.g., 3 active users on Lite).
Core features like custom roles require paid Business plan.
Flights (pipeline runtime) adds separate compute billing.

Technical limitations

Limited region availability (currently US East).
Relatively new ecosystem; some integrations may be less mature.

For technical folks

The deep-dive specs

FAQ

Includes technical Q&A

MotherDuck is the serverless cloud data warehouse built on DuckDB. It adds managed compute, shared storage, and a multi-tenant architecture (hypertenancy) on top of DuckDB's core engine, so you can scale from local dev to cloud production without infrastructure.