Legal Document Summarizer

The rising number of legal documents creates per- formance barriers for individuals who need quick information access since manual summary practices fall behind. The existing automated summary generation techniques utilize state-of-the- art NLP models for efficient text compression. This research puts forward a double-model approach which merges T5 for abstractive summarization coupled with BERT (BART-large- CNN) for extractive summarization applied to legal PDF texts after stopword filtering and content normalization. Experimental outcomes show the system achieves ROUGE-1 at 0.538 and ROUGE-2 at 0.250 along with ROUGE-L at 0.462 and BLEU at 0.038 and METEOR at 0.338 while using a reference legal summary for evaluation. The method simplifies the extraction of important details thus it improves the speed and accuracy of legal research together with case evaluation which establishes prospects for more complex document-intensive fields needing precise automated insights.