ASReview

About ASReview

ASReview is open-source AI-aided screening software for systematic reviews, coordinated at Utrecht University in the Netherlands. Systematic review screening is one of the clearest fits for active learning in all of research: a reviewer must read titles and abstracts for thousands of records to find the small fraction that are relevant, and the order in which those records are presented determines how much of that work is wasted. ASReview trains a model on the reviewer's own inclusion and exclusion decisions as they screen, continuously reordering the remaining records so relevant ones surface early. The project reports workload reductions of up to 95%, and the Financial Daily is quoted saying it compresses a year of work into two weeks. Three modes serve different users. Screen is the interactive reviewing interface. Simulate benchmarks model configurations against a completed review so you can test which classifier performs best on your kind of corpus before committing to it on a live one — a validation path that proprietary screening tools generally do not offer. Crowdscreen lets AI agents propose records to a panel of experts screening in parallel. Because it is open source and installable locally, records never leave the institution, which matters where corpora are licensed or confidential. Documented paths exist for screeners, information specialists, data scientists, developers, institutions and students.

Key Features

ASReview Pros & Cons

Tags