Practical guide · files stay on your device
Embedded assets, text and page pictures are different outputs.
A PDF page is not necessarily one embedded image. Extracting an original JPEG preserves that supported stream; rendering a page creates a new picture of its appearance. Text extraction reads an existing text layer, not text inside a scan.
Publisher and editorial responsibility: Project Hubbs · Updated
Open PDF asset inventory and extractionInspect your files for freeNo account required. Prepare resources explicitly before local processing. Inspection is free; cleaning and cleaned downloads require a license.
Supported input
- Supported PDF documents up to 25 MiB.
- Encrypted, signed, active or unsupported structures can be refused; do not weaken a document to bypass the check.
Key takeaways
- Inspect first for free. File format and structure—not the extension alone—determine eligible cleaning.
- An inventory and the selected supported original assets or existing text layer.
- Original-image extraction is limited to supported JPEG streams; other image filters are not renamed to PNG.
How to do it safely
- Choose a supported PDF in PDF studio and read its inventory before exporting.
- For original image bytes, export only the supported JPEG streams offered by the inventory. Read reasons for unsupported or masked images.
- For fonts or attachments, review the offered embedded programs and inert catalog attachments. A subset font is not a full font or a redistribution license.
- For searchable words, use text-layer extraction. A scan without usable text needs OCR, which PixelRinse does not provide.
- For a picture of an entire page, open PDF to images instead. Render up to 20 pages as PNG/JPEG/verified WebP, at 1200 px by default or optional 2400 px, within 32 MiB total.
- Review and download the outputs you actually need. Asset extraction does not clean metadata inside those extracted files.
What this workflow produces
- An inventory and the selected supported original assets or existing text layer.
- Page PNGs only when you separately choose page rendering.
What it does not promise
- Original-image extraction is limited to supported JPEG streams; other image filters are not renamed to PNG.
- Scans do not acquire a text layer.
- PDF page rendering and embedded-asset extraction are separate tools with separate output meanings.
Retained
- The original PDF.
- Unselected or unsupported resources remain in the original; extraction is not PDF sanitization.
Not checked
- OCR, faithful text-to-layout reconstruction or every unusual resource graph.
- Whether an extracted font or attachment is safe or licensed for another use.
Understand the evidence
Stream extraction, text extraction and page rasterization need different checks. A page PNG is a rendered output, not evidence that every embedded object was extracted.
Processing, comparisons and verification receipts stay local. PixelRinse sends no product-usage events. Payment and legacy license upgrades are separate online actions that never include selected file contents, names, hashes or metadata.
Common questions
More answers
Are PDF images and fonts extracted exactly as embedded?
Available original JPEG streams, embedded font programs and catalog attachments are exported without pretending they are rendered pages.
Technical details and limits
Unsupported filters, masked images, missing font programs and unusual resource graphs are inventoried with reasons. A font marked as a subset is not a full font, and extraction does not grant redistribution rights. Attachments remain inert and extraction does not clean metadata inside assets. Page-to-PNG rendering and text extraction are separate operations; scanned pages need OCR, which is unavailable.
Do document conversions retain the original page layout?
No. Document conversion creates text-only reflow, not a faithful copy of the original layout.
Technical details and limits
DOCX-to-PDF does not preserve tables, images, headers/footers or original pagination; its current font subset cannot encode every language. PDF-to-DOCX creates an editable text-only document, not a faithful reconstruction or OCR. This existing answer link is retained. The former video part now has its own answer: video comparison checks five synchronized samples, not every frame or audio.
What does Verified mean?
Verified means the bounded checks listed in the receipt passed on the rebuilt output.
Technical details and limits
Examples are that it parses, that supported metadata fields are gone and that retained content is intact. It describes those checks only, not every possible hidden trace.
Free inspection. One-time cleaning unlock.
Individual Pro is EUR 99 total, one-time, including applicable tax. No product tax is added on top. Lifetime (perpetual) use of the version you buy, plus 12 months of updates; update expiry does not revoke that access.
Merchant processing fees are covered. External network or issuer fees outside PixelRinse may be additional. Load purchase options on the pricing page to see currently configured methods.
Review pricing and payment options