A PostgresML deployment consists of multiple components working in concert to provide a complete Machine Learning platform. We provide a fully managed solution in our cloud.

* A PostgreSQL database, with pgml and pgvector extensions installed, including backups, metrics, logs, replicas and high availability configurations

* A PgCatpooler proxy to provide secure access and model load balancing across tens of thousands of clients

* A PgCatpooling proxy to provide secure access and model load balancing across tens of thousands of clients

* A web application to manage deployed models and host SQL notebooks

<figure><imgsrc="../.gitbook/assets/architecture.png"alt=""><figcaption></figcaption></figure>