<p>Cervical and ovarian cancers are among the most prevalent carcinomas affecting women worldwide, with over 70% of cases diagnosed at advanced stages, leading to poor survival rates and high mortality. Addressing the scarcity of well-annotated histopathological datasets, we present a novel benchmark dataset, the CervOvar-Cancer-Image-Vault (CIVa), specifically developed to support computational research and automated diagnostic tool development. CIVa comprises 1945 augmented histopathological images of cervical and ovarian lesions, systematically categorized into benign and malignant types to aid machine learning-based classification tasks. The images were captured using an Olympus CX23 microscope with a high-resolution camera from Hematoxylin and Eosin-stained whole-slide biopsy tissues, ensuring diagnostic clarity. Derived from samples of 67 patients at a multispeciality hospital, the dataset was carefully curated and verified by medical practitioners and pathologists to maintain clinical relevance. All images are standardized to 150 × 150 pixels and stored in “.jpg” format for easy integration into deep learning frameworks. To the best of our knowledge, CIVa is one of the first curated datasets focusing simultaneously on cervical and ovarian cancers, providing researchers with an authentic resource to develop, train, and validate advanced histopathological analysis models, ultimately contributing to early cancer screening and improved clinical decision-making.</p>