With the release of version 8.0, Elastic has introduced the capability to upload PyTorch machine learning models into Elasticsearch, enabling modern natural language processing (NLP) within the Elastic Stack. This integration facilitates the use of PyTorch models, which are popular for their support of deep neural networks like BERT, to perform a variety of NLP tasks such as sentiment analysis, named entity recognition, text classification, and text embeddings. Elasticsearch aims to provide a seamless user experience for uploading and managing these models with tools like the Eland client and Kibana's ML Model Management interface, while ensuring scalability and performance across clusters using the native libtorch library for inference. By incorporating NLP models directly into Elasticsearch, users can benefit from improved infrastructure, scalability, data security, and privacy, while also maintaining centralized management of models. The platform currently supports inference at ingest time and plans to expand to query time in the future, offering developers tools to build AI-powered search applications. Elastic encourages users to try out these capabilities with a free trial and provides resources for further learning and community engagement.