DuckDB
SQLite for analytics — in-process OLAP that queries Parquet, CSV, and S3 directly.
Acquired — still operating
AWS acquired DuckLabs Aug 2026
About
DuckDB is an in-process analytical database: a columnar OLAP engine that runs inside your application or notebook, speaks SQL, and queries Parquet, CSV, JSON, and cloud storage without a server, daemon, or cluster. Import the library and start running analytical queries on gigabytes of data with millisecond startup.
2025 was DuckDB's breakout year — 1M+ downloads per day and 40M+ monthly — and in August 2026 AWS acquired DuckLabs, the Amsterdam company employing the core team. The project itself stays MIT-licensed under the independent DuckDB Foundation, with creators Hannes Mühleisen and Mark Raasveldt continuing to lead it. AWS has since embedded DuckDB inside Aurora PostgreSQL and uses it across Quick and its agentic-AI data services.
Limitations: it is deliberately single-node — not a replacement for distributed warehouses on truly massive data; concurrent write support is limited compared to transactional databases; and while the Foundation structure protects the license, the AWS acquisition means watching governance for any future license shifts.
What it can do
- In-process columnar OLAP engine
- Direct queries over Parquet, CSV, JSON, S3
- MIT-licensed and embeddable
- MotherDuck cloud collaboration layer
- WASM builds for browser analytics
Setup
pip install duckdb (or the CLI/binary for your platform); import duckdb and query files directly — no server to start.
Prerequisites
Python, R, or a shell; data files or object storage to query.
Use cases
Works with
Trust score — 9/10
Scored 9/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 67; Trending score 90/100 from researcher assessment.
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.