A fork of PostgreSQL

for real-time machine learning

How VidarDB Works?

VidarDB can do everything PostgreSQL can do, but with our technology, a polymorphic storage engine, we make VidarDB run like a speedy antelope.

PostgreSQL Frontend

Because VidarDB embraces PostgreSQL ecosystem, it uses the same terminal and driver as PostgreSQL. No need to change your code, while all the miracles happen underneath.

In-place Machine Learning

It is never required to dump your data from the database to your machine learning tool anymore. VidarDB simplifies your process by running your machine learning algorithm directly in the database via your familiar SQL.

Blazingly Fast Querying

Unlike general operational database, VidarDB is designed for modern memory hierarchy, with up to orders of magnitude faster in data ingestion and querying recent data. Fresh data sit in RAM first, and will be automatically flushed to secondary memory when RAM is full.

Responsive Data Analysis

VidarDB natively supports business intelligence. The Column Store, an integral part of our innovative data structure, enables data analysis extremely responsive by fetching only portion of the data. Unlike general data warehouse, VidarDB ingests data up to orders of magnitude faster.

Typical Examples

Linear Regression

Logistic Regression


Stay in the loop.

Sign up now to get priority notifications about VidarDB.

Supported by