Image
Image de-identification is coming soon.
This page is a placeholder for future Guardian Layer support for images, screenshots, scanned documents, and other visual inputs that may contain sensitive information.
Current Status
Guardian Layer currently supports text and document workflows, including:
- Text input.
- CSV, XLSX, and XLS files.
- DOCX files.
- PDF files with extractable text.
- TXT files.
Image-specific de-identification is not currently exposed as a public API workflow in this documentation set.
What Image Support Is Expected to Cover
When available, image de-identification may be useful for inputs such as:
- Screenshots that contain names, email addresses, phone numbers, or account details.
- Scanned forms.
- ID-like documents.
- Images embedded with visible sensitive text.
- Visual records that need privacy review before sharing or storage.
The exact supported image formats, request parameters, and output behavior should be confirmed when the image API is released.
What to Use Today
If your content is not image-only, use the existing Guardian Layer workflows:
| Input type | Use today |
|---|---|
| Plain text | Text de-identification |
| CSV, XLSX, or XLS | File De-identification |
| DOCX | File De-identification |
| PDF with extractable text | File De-identification |
| TXT | File De-identification |
| Scanned image-only PDF | Not covered by the current file workflow |
| PNG, JPG, or screenshot | Coming soon |
For supported file workflows, see File De-identification.
Product Note
Image de-identification usually requires additional steps that are different from text-only processing, such as text extraction, layout detection, visual masking, and output rendering.
Because of that, image support should have its own request examples, limitations, and quality guidance once the feature is available.
Before This Feature Is Available
If users need to process image-only content now, clarify whether they can:
- Convert the content into extractable text.
- Use a PDF or document version that contains selectable text.
- Manually review the image before sharing.
- Wait for the dedicated image workflow.
Avoid presenting image de-identification as available until the API and Dashboard flow are confirmed.