ASReview

Systematic literature reviews require researchers to screen thousands of academic titles and abstracts, a process that traditionally consumes hundreds of hours. Developed at Utrecht University, ASReview automates this pipeline using active learning algorithms that prioritize relevant publications based on user feedback. By continuously retraining its machine learning model during the screening session, the tool allows researchers to locate up to 95 percent of relevant studies while reading only a fraction of the total dataset. Built as an open-source framework, ASReview guarantees full algorithmic transparency, repeatable research outputs, and strict compliance with global data privacy standards. The platform supports multiple text classification models, active learning strategies, and customizable feature extractors, making it suitable for screeners, information specialists, and data analysts across diverse domains. Furthermore, its crowd-screening functionality enables collaborative team reviews to accelerate large-scale research projects. While the software relies on user input for iterative learning and lacks native multi-language screening capabilities, ASReview serves as an essential utility for academic institutions and policy agencies worldwide.