DETECTING DOCUMENT TEXT THAT IS HARD TO READ

A computer system is configured to determine portions of text extracted from a corresponding group of documents; process a particular portion of text by a set of filters, where the particular portion of text may correspond to a particular document, and where each of the filters may generate a respective score based on processing the particular portion of text; calculate a readability score based on the respective scores generated by the filters; determine that the readability score satisfies a threshold score; and generate or select a new portion of text, for the particular document, based on determining that the readability score satisfies the threshold score.