---
title: GLiNER Streaming PII Detector
emoji: 🔍
colorFrom: gray
colorTo: indigo
sdk: gradio
sdk_version: 6.20.0
app_file: app.py
short_description: Zero-shot PII detection with GLiNER streaming-span model
python_version: "3.12"
startup_duration_timeout: 30m
---

# GLiNER Streaming PII Detector

This Space demonstrates [knowledgator/gliner-stream-pii-v1.0](https://huggingface.co/knowledgator/gliner-stream-pii-v1.0),
a 0.6B-parameter zero-shot PII/NER model built on a Qwen3 backbone using the
GLiNER streaming-span architecture.

## How it works

- Paste any text that may contain personally identifiable information (PII).
- Specify the entity types you want to detect (person, email, phone number, etc.).
- Adjust the confidence threshold to control precision vs. recall.
- The model highlights detected entities directly in the text and lists them in a table.

The model is **open-label** — you can use any entity type name, not just the defaults.