fastino/gliner2-privacy-filter-PII-multi is a GLiNER2 model specialized for token classification in named entity recognition tasks focused on personally identifiable information. Hosted on Hugging Face, it addresses the need to detect and handle sensitive data in text across multiple languages.
The model supports extraction of 42 types of PII labels and operates in seven languages. It is delivered as a Safetensors-format model under the apache-2.0 license and is trained on synthetic data. Integration occurs through the GLiNER2 library with a simple from_pretrained call, after which users supply text and a list of entity types such as company, person, product or location to receive extracted spans.
Documentation on the model card includes instructions for notebooks on Google Colab and Kaggle, benchmark results on the SPY dataset, guidance on when to apply the model, and an example of redaction. It draws on two arXiv papers for its underlying approach. The model card positions it within the class of span-extraction and information-extraction tools for privacy and redaction workflows.
Gliner2 Privacy Filter PII Multi sits in PulseGate's Other AI category. Automatically identifying and masking sensitive personal information in text to ensure privacy compliance. It is built as an open-source project for developers. Gliner2 Privacy Filter PII Multi is open source under the Apache-2.0 license. Gliner2 Privacy Filter PII Multi is available on the web, the command line, and API.
Behind Gliner2 Privacy Filter PII Multi is fastino, and the product first shipped in 2025. Development happens publicly on GitHub with 1.7k stars and 23 commits in the last 90 days. Key capabilities include PII Detection, Privacy Redaction, and multilingual.
Latest indexed changes and source events
fastino/gliner2-privacy-filter-PII-multi verified by the PulseGate indexer
Other apps tracked under the same category.
