True PDF redaction is fundamentally different from visual annotation. A common mistake is to use a PDF annotation tool to place a black rectangle over sensitive text — this merely overlays the content, which remains in the file and can be extracted by removing the annotation layer. Proper redaction, as implemented in this tool, works by drawing new vector shapes (filled rectangles) directly into the PDF page's content stream using pdf-lib. These shapes physically overwrite the coordinates where the sensitive content existed. When the redacted PDF is saved, the original text or image data at those coordinates is replaced by the solid fill, making recovery impossible. The interactive drawing interface renders each PDF page using pdf.js, providing a visual canvas where you can click and drag to define redaction areas. The page rendering is high-fidelity, ensuring that you can precisely target the content that needs redaction without accidentally covering adjacent text.