Action ML Logo

Universal Recommender

The Universal Recommender (UR) uses almost anything known about the user to make recommendations better. It takes in user actions and profiles, context, and item metadata, all of which improves quality. It is one of a new generation of multimodal recommenders that have obsoleted most of the older technology available. The UR is perhaps the only publicly documented end-to-end recommender to do this.

Launch on AWS


Set up a free consultation

Make use of any part of user's clickstream or even profile and context information in making recommendations.

Multiple data icon

Use all your data

The Universal Recommender can take in user actions, profile data, contextual information, virtually a user's entire clickstream, add that to content and properties to gain deep understanding of user preferences, others cannot.

Flexibility icon


The Universal Recommender has extremely flexible but simple queries for most any situation. It is also open source built so you can choose the hosting environment that works for you—start simple or customize any way that suits you.

Data and model icon

Own your data and model

Your data is your competitive advantage. With The Universal Recommender, you keep it private. Some SaaS vendors see this differently—not us. SaaS providers also charge for their operations—something you may not need.

Lock-in icon

No lock-in

It’s open source and no vendor lock-in. You can customize the code to fit your needs at anytime. You can't do that with AWS, Google, or Azure services.

Practical solutions, powerful examples.