TextAttack is a Python framework for adversarial attacks, data augmentation and model training in natural language processing. Security and machine-learning teams use it to test how text classifiers and other NLP models behave when inputs are deliberately altered, to generate additional training data, and to train models on adversarial examples. It is a library that you install with pip and run locally or inside your own pipelines. There is no vendor-hosted service, and the code is released under the MIT licence.
The project is maintained by the QData Lab at the University of Virginia in the United States and is published on GitHub. Because it runs in your own environment, the models and texts you test are not sent to the maintainers, and no hosting region applies. The maintainers are an academic research group rather than a company, so there is no commercial support contract or service-level agreement. For a European organisation it is a self-hosted component for robustness testing of NLP systems, comparable in purpose to commercial adversarial-testing tools offered as cloud services.