About DuckDB
DuckDB is an analytical SQL database that runs inside your program, so there is no separate server. It is built for fast queries over large tables, such as reports and data science work.
What it does
You can query local files or data in object storage with SQL. DuckDB reads common formats including CSV, JSON, Parquet and Iceberg, and it can write several of them too. Its SQL dialect is inspired by PostgreSQL, so many queries will feel familiar.
Key features
- Runs in-process with no server
- Queries CSV, JSON and Parquet files directly
- Clients for major programming languages
- Extensions for new formats and functions
- Works with data in object storage
Who it's for
Data analysts who work with large files, Python and data science users, and developers who want an embedded analytics engine in a desktop or edge app.
Pricing and good to know
DuckDB is free and MIT licensed, and the project is governed by an independent foundation. It runs on laptops, servers and edge devices. Because it is built for analytics, it is not meant to replace a transactional database that handles many small writes from users at the same time.
Highlights
- Runs inside your app with no server to manage
- Queries CSV, JSON and Parquet files with SQL
- Clients for Python, Node.js, Go, Java and Rust
- Reads Iceberg and Delta Lake formats
- Extensions that add functions and data sources
