A smaller PDF is not always a better PDF. Some compression methods reduce image data while keeping the original text layer; others redraw every page as an image. That second approach can make a file smaller, but text that you could previously select or search may disappear. The right choice depends on whether your document is a digitally created PDF or a scan, and whether search, copying, links, forms or signatures need to keep working.
Key takeaways
- First check whether the original PDF contains selectable text; a page that only looks like text may be a scanned image.
- SortMyFile Compress PDF rebuilds each page as a JPEG image, so it does not preserve the original searchable text, links, forms or annotations.
- If a scan needs search, OCR can add a new English text layer, but it does not restore the original document structure and recognition must be checked.
- Keep an untouched copy and compare readability, page count and important details before uploading the result.
Why PDF compression can remove searchable text
A PDF can contain actual text objects, vector graphics, images and interactive elements. In a digitally created document, words may be stored as text that a viewer can select, copy and search. In a scanned document, the page may be one large image; it can look like printed text while containing no text objects at all. A compression workflow that renders every page to a bitmap and places that image into a new PDF replaces the original page content with pixels. The page may still look readable, but the original text layer, vector detail, links, form fields and annotations are not carried across.
Check whether your PDF is searchable before changing it
Open the original in a PDF viewer and try to select a sentence. Then use the viewer's Find command to search for a word that is clearly visible on the page. If both work, the document likely has a usable text layer. If you can only select the whole page as an image, or Find returns nothing, it may be a scan without OCR. A few pages can differ: a document may contain digitally generated text on some pages and scanned images on others. Check representative pages rather than relying on the cover page alone.
- Open the original PDF in a trusted viewer.
- Drag across a sentence and see whether individual words can be selected.
- Use Find to search for a word printed on a different page.
- Repeat on pages that contain scans, tables, stamps or inserted images.
- Keep the untouched original before testing any compression output.
Choose a compression method that matches the document
If the PDF contains searchable text, links, forms or annotations that must remain usable, do not use a workflow that rasterizes every page. Look for a document-aware optimizer that reduces embedded image resolution or removes redundant objects while retaining text and interactive structures, then verify its output. If the file is already a scan and the only requirement is a smaller visual copy, image-based compression may be acceptable, provided small print and signatures remain legible. The receiving portal's file-size limit does not override the need to preserve information.
What the SortMyFile compressor preserves—and what it changes
SortMyFile Compress PDF renders each page in the browser and rebuilds the output using JPEG images. Its Low, Recommended and Extreme presets use different render scales and JPEG quality settings; Target size (KB) tests up to six quality settings and returns the best candidate it found. This is an image-based workflow, not a text-preserving PDF optimizer. The exported pages do not retain the original selectable text, vector graphics, hyperlinks, form fields, annotations or other interactive page content. Fine print, QR codes, stamps and signatures can also become softer. The requested KB target is not guaranteed: if no tested result fits, the smallest tested output is returned even when it remains above the target.
Can OCR make a compressed scan searchable again?
If you have an image-based PDF and need search, OCR can recognise printed characters and add a text layer. SortMyFile OCR PDF currently recognises printed English. It renders the pages and creates a new image-based PDF with an invisible OCR text layer; it does not restore the original vector content, links, form fields, bookmarks, annotations, tags or metadata. Recognition can confuse similar letters and digits, especially in low-resolution scans, columns, handwriting or faint print. Proofread names, dates, account numbers, totals and other important values against the page image. OCR can make text searchable, but it is not the same as preserving the original text objects or document accessibility structure.
A safer workflow for reducing PDF size
Use the following decision process before you upload a smaller copy. If the document has a text layer or interactive content that matters, preserve the source and use a text-preserving optimizer rather than raster compression. If it is a scan, decide whether search is required before compressing. When OCR is needed, test the result carefully and keep both the original scan and the final searchable copy.
- Make a separate working copy; never overwrite the only original.
- Test text selection and Find on several pages.
- List anything that must survive: selectable text, links, form fields, annotations, signatures, vector diagrams or bookmarks.
- If those features matter, choose a text-preserving optimization method and verify the output.
- If the source is a scan and a smaller image-based PDF is acceptable, use a conservative setting first and inspect the result.
- If the compressed scan must be searchable, run English OCR and proofread the recognised text.
- Compare page count, visual clarity and required document features before submitting.
What to inspect before uploading the compressed PDF
Open the downloaded file in a separate viewer. Check the first page, a dense page, a page with a signature or stamp, and the final page. Try selecting and searching text again if searchability is a requirement. Open links and form fields if they are part of the workflow, and inspect QR codes at normal zoom rather than assuming they still scan. Compare the output size with the portal's stated limit, but do not submit a file that meets the number while making essential information unreadable. Keep the original until the recipient confirms acceptance.
Put the guide into practice
Open the related tool when it fits your task, and check its supported formats and limitations before downloading or submitting a file.