Model Setup
Upload your BERT NER model components to get started. All files stay on your device.
ONNX Model File Required
Upload your BERT Base NER model (.onnx). Works with dslim/bert-base-NER and other BERT NER models.
Tokenizer Config Optional
Upload tokenizer.json for custom tokenizer settings. Default BERT base uncased config is used otherwise.
Vocabulary File Optional
Default BERT base uncased vocab (30,522 tokens) is loaded automatically. Upload custom vocab.txt to override.
optimum-cli export onnx or transformers.onnx.
Process Document
Upload a PDF to extract and analyze personally identifiable information.
Drop your PDF here
or click to browse
Review Entities
Review detected entities and toggle which ones to redact. Click any row to highlight it in the PDF preview.
| Redact | Original Text | Type | Page | Confidence |
|---|
Finalize Redaction
Load a PDF and run analysis, then click an entity to see it here.
Restore Document
Upload a redacted PDF and its encrypted key file to restore the original text.
Redacted PDF
Upload the previously redacted PDF
Encrypted Key File
Upload the .json key file generated during redaction