PostgreSQL

Picking a database for a new project often means compromise. MySQL is fine until you hit complex queries, JSON fields or geographic data, and commercial databases have everything but come with license fees and vendor lock-in.

PostgreSQL is where many teams end up. It's an open-source object-relational database with more than thirty years behind it, strict about the SQL standard and reliable with transactions, yet able to store JSON like a document database and pick up geographic data (PostGIS), full-text search and vector search through extensions. A lot of the apps and sites you use every day run on it.

It's completely free and open source under a permissive license, with official installers for Windows, macOS and Linux.

Features

Rich data types: beyond numbers, text and dates, native JSON/JSONB, arrays, key-value pairs and geometric types, plus custom types.

Reliable transactions: full ACID transactions and MVCC so reads and writes don't block each other, with constraints that keep out duplicates and nulls.

Advanced queries: window functions, common table expressions, recursive queries and materialized views, so complex reports stay in SQL.

Many index types: B-tree, hash, GiST, GIN, BRIN and more, chosen to fit the data, speeding up full-text and JSON queries.

Extensions: PostGIS for geographic data, pgvector for vector search and more, plus stored procedures in several languages.

Replication and HA: built-in streaming and logical replication for read replicas and failover.