MLDB is proud to have been acquired by ElementAI to build upon the innovative work already being conducted on behalf of our many clients. Joining forces with ElementAI will enable us to bring even more experience and expertise to our current projects and we look forward to the next chapter of world-changing innovation.

MLDB is an open-source database designed for machine learning.
You can install it wherever you want and send it commands over a RESTful API to
store data, explore it using SQL, then train machine learning models and expose them as APIs.

Train models faster than H20, Spark MLlib or scikit-learn. Then call them over HTTP thousands of times per second.

Vertical scaling means you can use all the RAM and all the cores. Process billions of data points on machines that cost $1/hr.

The Community Edition is available on Github under the Apache license. Come chat with us on Gitter!

SQL is still the best way to query and manipulate data. But with support for millions of columns, good thing it's schema-free.

Supervised or unsupervised, numbers or images or text, we've got it covered. We also support TensorFlow graphs.

Installation with Docker, Data Science with Jupyter. Read and write text, CSV or JSON to & from local storage or directly on S3 or HDFS.

Uniform JSON-over-REST API, accessible from any language. Comes with a nice Python wrapper called pymldb.

As soon as you train a model, it's available as an HTTP endpoint. No deployment necessary.