MOTIVATION: The Adverse Outcome Pathways (AOP) framework advances alternative toxicology by prioritizing the mechanisms underlying toxic effects. It organizes existing knowledge in a structured way, tracing the progression from the initial perturbation of a molecular event, caused by various stressors, through key events across different biological levels, ultimately leading to adverse outcomes that affect human health and ecosystems. However, the increasing volume of toxicological data presents a significant challenge for integrating all available knowledge effectively. RESULTS: Text mining techniques, including natural language processing and graph-based approaches, provide powerful methods to analyze and integrate large, heterogeneous data sources. Within this framework, the AOP-helpFinder TM tool, accessible as a web server, was created to identify stressor-event and event-event relationships by automatically screening scientific literature in the PubMed database, facilitating the development of AOPs. The proposed new version introduces enhanced functionality by incorporating additional data sources, automatically annotating events from the literature with toxicological database information in a systems biology context. Users can now visualize results as interactive networks directly on the web server. With these advancements, AOP-helpFinder 3.0 offers a robust solution for integrative and predictive toxicology, as demonstrated in a case study exploring toxicological mechanisms associated with radon exposure. AVAILABILITY AND IMPLEMENTATION: AOP-helpFinder is available at https://aop-helpfinder-v3.u-paris-sciences.fr.