Data Warehouse Built for Agents
Fast, serverless analytics powered by DuckDB — with dedicated compute for every user and every agent. Build a data stack in one prompt, or plug into the one you have.
Or start with your coding agent:
Trusted by builders
Which duck species flies furthest in fall migration?
An Agent-First Architecture
Hypertenancy: an isolated compute instance for every person, dashboard, and agent. Run fast, highly concurrent queries without resource contention.
Every person, dashboard, and agent session gets its own isolated DuckDB instance — a Duckling — spun up in milliseconds.
Agents fan out — every session gets its own isolated duckling, so a fleet of parallel agents never slows your dashboards. Cold starts land under 100ms.
One user's heavy query saturates their duckling — and nobody else's. No noisy neighbors, no queues, no warehouse-wide slowdowns.
Size each instance to its job, Pulse through Giga. Backfill on a Mega instance, run dbt workloads on a Standard, let end users query a fleet of Pulses.
Shared Storage
DuckLakeDifferential StorageApache IcebergObject Storage
Serverless Analytics for Teams at the Frontier
Everything between your data and an answer — compute, storage, and the agent tooling that ties them together.
Work with the tools you know
Build with the tools you love, from the latest in the data stack to your favorite languages.
Integrations & Tools
Choose your stack
Customers building on MotherDuck
Production apps, customer-facing analytics, and internal insights — shipped on MotherDuck.
Agent-native Analytics
Agents don't query like humans do. Use the skills, tools, and interfaces that fit your workflow best.
Claude Code v2.1.220
Opus 5 (1M context) with xhigh effort · Claude Team
~/analytics
▶▶ auto mode on (shift+tab to cycle) · esc to interrupt
Any MCP-capable agent
Zed, ChatGPT, Claude Desktop, and more
Fresh from the flock
Tips, tricks, product news, and helpful resources from the MotherDuck team and beyond.