The iRead4Skills is an open access intelligent system that evaluates texts complexity.
It assists trainers and content producers in creating or adapting texts with the appropriate level of complexity and can also to be used to evaluate and suggest readings according to the user literacy level.
Text complexity is measured based on diverse text features and does not reflect the value or literary quality of the content.
The system analyses texts only. It is not designed to grade or assess individuals.
Supported languages: French, Portuguese and Spanish.
Overview
The iRead4Skills system was developed to support and motivate low-literacy individuals in improving their reading skills by helping them select reading materials that interest them, thereby facilitating their access to information and culture.
The system was designed in close cooperation of low-literacy adults and professionals in Adult Learning and Vocational Education. It analyses the complexity of texts automatically using supervised machine learning, based on a set of texts carefully selected and manually classified by experts into three complexity levels specifically tailored to the target audience. The development of the tools was supported by end-user evaluation and validation methods, and complemented by expert validation from trainers experienced in working with low-literacy adults, ensuring consistency throughout.
The analysis provided by the iRead4Skills system is transparent and consistent. It considers several complexity dimensions and features, displayed in the interface, which collectively contribute to the overall assessment of a text. These multidimensional indicators should not be interpreted in isolation or used as the sole basis for analysis, as complexity assessment is inherently subjective and may be perceived differently by different individuals. Text complexity is measured solely through linguistic features and does not reflect the value or literary quality of the content. Books and other longer texts may vary in complexity across different sections.
As an automated system, the iRead4Skills tools produce systematic and consistent results. However, while the system provides reliable outputs, these are not free from error or subjectivity and should be interpreted as informed indicators and not absolute measures.
The system’s suggestions for adapting texts take into account the multiple aspects that jointly determine text complexity. These suggestions can help users adjust a text to a specific target level, but other revision strategies are equally valid. Users remain in full control of how the text is edited and may accept or reject suggestions at any time.
No text is stored during the process: all texts or text images are discarded after analysis.
The iRead4Skills system is intended to help teachers and trainers select relevant reading materials for low-literacy adults and to motivate and empower individuals to improve their reading skills and access information and culture. It does not compete with publishers’ offerings, but rather, it provides a mechanism for structured complexity-based classification of works, supporting discoverability, reader confidence, and literacy progression. In practice, this contributes to widening readership, particularly among adults with emerging or intermediate reading skills, thereby expanding the potential market for books rather than diminishing it. The system can also assist content producers in adapting texts for low-literacy audiences or to ensuring that information is accessible to a wider range of readers.
