AI Prep Tools

Hidden Text & Invisible Character Scanner

Locate invisible Unicode characters and hidden document markup locally before feeding files to AI.

  1. Paste text or drop a .txt, .md, .html, .docx, or text-layer .pdf file.
  2. Inspect every located finding with page/offset context and why it triggered.
  3. Export a cleaned plain-text copy stripped of invisible Unicode controls.
Loading local browser workspace...

Supported Formats, Limits & Accuracy

Unicode and DOM/DOCX/PDF rule detections are Exact for the inspected patterns; overall evasion coverage is incomplete and optional instruction heuristics are Heuristic.

Supported:
Text, HTML, DOCX, Text-layer PDF
Configured limit:
25 MB (.txt, .md, .html, .docx, text-layer .pdf)
Processing:
Client-side browser execution ( how to verify)

Honest Limitations

  • Some evasion techniques cannot be detected by static inspection; a 'No known patterns found' result is never a guarantee that a document is harmless.
  • Uploaded HTML and embedded PDF scripts are inspected as raw structure and are never executed or rendered as active DOM.
  • Scanned image-only PDFs have no text layer to inspect.

How to Use Hidden Text & Invisible Character Scanner

  1. Paste text or drop a .txt, .md, .html, .docx, or text-layer .pdf file.
  2. Inspect every located finding with page/offset context and why it triggered.
  3. Export a cleaned plain-text copy stripped of invisible Unicode controls.

Frequently Asked Questions

Why doesn't the scanner display a green 'Safe' badge when 0 findings appear?
Because determined prompt-injection concealment (such as text drawn inside rasterized images) can bypass structural scanners. We report 'No known patterns found' so you always treat untrusted documents carefully.

Next Steps & Related Tools