Skip to content


Repository files navigation



Recotem is an easy to use interface to recommender systems; Recotem can be launched on any platform with Docker. It ships with a Web-base UI, and you can train and (qualitatively) evaluate the recommendation engine solely using UI.

Sample usage of recotem

Recotem is licensed under Apache 2.0.



Getting Started

There are two ways to start using Recotem. Both requires latest docker.

1. Using pre-built image.

  1. Visit latest release
  2. Download "Docker resources to try out" from Assets
  3. Unzip it and
    • (Windows) Click "recotem-compose" script
    • (Linux & MacOS) Run docker-compose there.
         docker-compose up`

See for a friendlier introduction.

2. Building the image

  1. Clone this repository.
  2. In the repository top directory, simply run
        docker-compose up


Backend & Worker

To run the backend (and worker) in Django development mode, use docker-compose-dev.yml.

docker-compose -f docker-compose-dev.yml build
docker-compose -f docker-compose-dev.yml up


To run the frontend webpack-dev-sever, you will need a descent version of yarn.

After yarn under frontend/ directory to install the dependency, run

cd frontend
yarn serve

In order for the frontend to work with the API, you first have to launch the backend following the above instruction.

Command-line tool

recotem-cli allows you to

  • tune & train recommender systems
  • obtain the recommendation result

via command-line interface.

Batch execution on ECS

There is an example project which uses recotem to batch-execute recommendation task on Amazon ECS.