Compress a PDF Online Without Uploading It [2026]
Every popular PDF compressor uploads your file to a server. SammaPix is different: it compresses PDFs entirely in your browser, with no upload, no signup, and no server involved. Here is exactly how it works, what it compresses well, and where it is honest about its limits.
Table of Contents
The problem: every PDF compressor uploads your file
You have a 25 MB scanned invoice. Your client expects a PDF under 5 MB. You search for "compress PDF online" and land on iLovePDF, Smallpdf, or Adobe Acrobat Online. You drag the file in. A progress bar runs. The file uploads to their server, gets processed remotely, and you download the result.
The compression worked. But your invoice, which may contain payment terms, client names, amounts, and your account details, spent time on a server you have no control over. Their privacy policy says they delete it after an hour. You have no way to verify that.
For photos and generic content this is a minor concern. For scanned contracts, tax documents, medical bills, and financial records, it is a real risk. I built SammaPix PDF Compress to eliminate that risk by running the entire compression process inside your browser. No server is involved at any point.
What "no upload" actually means
When you drop a PDF into SammaPix PDF Compress, your browser reads the file using the FileReader API — a standard browser API for reading local files without network access. The file stays in browser memory. Processing happens using pdf.js (Mozilla's open-source PDF renderer) and pdf-lib (an open-source JavaScript library for reading and writing PDFs). The output is served as a blob: URL and downloaded directly from memory. Zero network requests carry your file.
How browser-based PDF compression actually works
Understanding the mechanism explains both the strengths and the honest limitations of this approach. Here is what happens under the hood when you click Compress:
- pdf.js renders each page to a canvas. Each page of your PDF is drawn pixel by pixel onto an HTML canvas element, just like a browser renders a PDF for viewing. The render resolution is controlled — lower quality settings use a lower pixel density.
- Each canvas is exported as a JPEG. The canvas is converted to a JPEG image at the quality level you selected (Low, Medium, or High corresponds to different JPEG quality values). JPEG compression is lossy but produces very compact files for photographic content.
- pdf-lib assembles a new PDF. Each JPEG is embedded as a full-page image in a new PDF document. The output PDF contains one image per page.
- The result is offered for download. The new PDF is stored as a Blob in browser memory and downloaded directly. Nothing leaves your device.
This is called rasterization: the PDF content is converted to a raster image (pixels) rather than preserved as vectors and text. This is the same technique used by scanners when they save a physical document as a PDF. It is extremely effective at reducing file size for image-heavy PDFs, and comes with one important trade-off explained next.
What this tool compresses well — and where it is honest
Not all PDFs are the same internally. The compression results vary significantly based on what the PDF actually contains.
| PDF type | Typical size reduction | Why |
|---|---|---|
| Scanned documents (invoices, forms, receipts) | 50 to 80 percent | Already image-based. Re-encoding at lower JPEG quality removes a lot of data that the eye cannot see. |
| Photo portfolios, brochures, catalogues | 40 to 75 percent | High-resolution embedded images compress well. Good for sharing, less good for print. |
| Presentations exported to PDF | 30 to 60 percent | Mixed content. Slide backgrounds and embedded photos compress well. Charts and text less so. |
| Text-only PDFs (reports, legal documents, ebooks) | 0 to 20 percent (may increase) | Vector text compresses efficiently in the PDF format. Rasterizing it to JPEG is not efficient for sharp edges and small type. The tool detects this and keeps the original rather than making it larger. |
| Already compressed PDFs | Minimal | If the file was already run through iLovePDF or Acrobat compression, the images inside are already encoded at a lower quality. Re-compressing gives diminishing returns. |
The honest limitation to understand: because this tool rasterizes pages, the text in the compressed PDF becomes part of the image. It will not be selectable or searchable in the output. For most sharing and archiving use cases this is fine. For legal or contractual documents where you need recipients to select, copy, or search the text, a server-side compressor using font subsetting (such as Ghostscript) is the right tool.
Compress your PDF in your browser now
Ideal for scanned documents, invoices, and image-heavy PDFs. See the before and after size before downloading. No upload. No signup. Free.
Open PDF Compress, FreeHow to compress a PDF without uploading it, step by step
The full process takes under a minute for most PDFs:
- Go to sammapix.com/tools/pdf-compress in Chrome, Safari, Firefox, or Edge. No account required.
- Drop your PDF onto the dropzone or click to browse for it. The file is loaded into browser memory. You will see the original file size displayed.
- Select a quality level: Low, Medium, or High. Medium is the recommended starting point. See the next section for guidance on which to choose.
- Click Compress. pdf.js renders each page and pdf-lib assembles the output PDF. The before and after file size are shown. Processing takes 5 to 30 seconds depending on the number of pages and your device speed.
- Review the result. The tool shows you the compressed file size before you download. If the reduction is not enough, try the Low quality setting for a more aggressive result.
- Download the compressed PDF. Click Download. The file is served from browser memory as a Blob. No network request occurs.
Choosing the right quality level: Low, Medium, or High
The three quality levels control the JPEG encoding quality applied to each rasterized page. Here is how they differ in practice:
Low — smallest file, visible quality loss
Low produces the most aggressive compression. Text remains readable, but images may show JPEG compression artifacts (blocking, banding) when viewed close up. The file size reduction is typically 65 to 80 percent for scanned documents. Use Low when the goal is to get the file under an email attachment limit and the recipient does not need high-fidelity images.
Medium — balanced (recommended for most use cases)
Medium reduces file size by 50 to 70 percent while keeping pages visually clean at screen resolution. Artifacts are minimal. Text is clear. Images look good at normal viewing distance. This is the right choice for sharing invoices, receipts, and scan bundles with clients or colleagues.
High — near-original quality, moderate reduction
High preserves image quality close to the original. Visible difference from the source is minimal even when zooming in. File size reduction is typically 20 to 40 percent. Use High when the compressed PDF will be printed or when image fidelity matters, such as a photo portfolio or a brochure that a client will review closely.
| Quality level | Typical size reduction | Best for |
|---|---|---|
| Low | 65 to 80 percent | Archiving, getting under email size limits, internal sharing where print quality is not needed |
| Medium | 50 to 70 percent | Client-facing invoices, receipt bundles, most email attachments, scan-heavy documents |
| High | 20 to 40 percent | Brochures, portfolios, photo-rich PDFs, files that will be reviewed at high zoom or printed |
Browser-based vs upload-based PDF compressors: honest comparison
Here is an objective comparison across the dimensions that matter for someone choosing a PDF compressor:
| Dimension | Upload-based (iLovePDF, Smallpdf, Adobe Online) | Browser-based (SammaPix) |
|---|---|---|
| Privacy | File uploaded to a remote server. You trust their deletion and security policies. | File never leaves your device. Verifiable via browser Network inspector. |
| Text selectability after compression | Preserved if the tool uses font subsetting (Ghostscript-based). Not preserved if it rasterizes. | Not preserved. Pages are rasterized to JPEG. Text in the output is part of the image. |
| Size reduction for scanned PDFs | Excellent. Server tools can also optimize streams and embedded objects. | Excellent. JPEG re-encoding of image-based pages is very effective. |
| Size reduction for text-only PDFs | Better. Server tools can subset fonts and compress text streams without rasterizing. | Limited. Rasterizing vector text is not efficient. The original is kept if the output would be larger. |
| File size limits | Free plans typically cap at 5 to 25 MB. Larger files require a subscription. | Limited by device memory. No artificial cap. Very large PDFs may be slow on low-RAM devices. |
| Speed | Depends on upload speed and server load. Can be slow on large files. | Depends on your device CPU. Fast on modern hardware. No network latency. |
| Signup required | Often required beyond basic free use. | No account required. |
| Works offline | No. | After the page loads, compression works without an internet connection. |
How to verify no upload happens
You do not need to take my word for it. Here is how to verify this yourself in under two minutes:
- Open DevTools. Press F12 (Windows/Linux) or Command Option I (Mac) in your browser. On Safari you may need to enable the Develop menu first in Settings → Advanced.
- Go to the Network tab. Click on the Network panel in DevTools. If you see a list of requests, click the clear button to empty it.
- Drop your PDF and click Compress. Watch the Network panel as the tool processes the file.
- Observe: no outgoing requests. You will see no network activity during compression. The only requests that appear are the initial page load assets. Nothing carries your file to any server.
This is the same verification method that security researchers use to audit tools that claim to be privacy-safe. It is straightforward and conclusive.
Your PDF stays on your device
No upload, no account, no server. Works on scanned invoices, photo PDFs, and brochures. Keeps the original if the output would be larger. Free.
When to use a different tool
Browser-based rasterization compression is excellent for some use cases and the wrong choice for others. Here is when to use something else:
You need the text to remain selectable after compression
If the PDF is a legal document, a signed contract, or a report where recipients need to select and copy text, use a Ghostscript-based tool instead. Ghostscript compresses PDFs by subsetting fonts and optimizing streams while preserving the text data. It runs from the command line or is available via server-based tools. The privacy trade-off is that it requires either installing software locally or uploading to a service.
The PDF is already well-compressed
If the file was already processed by iLovePDF or a similar tool, the internal images are already at a lower quality. Re-compressing will not help much. The tool shows you the size before and after, so if the reduction is less than 10 percent, the file is already well-optimized. Consider splitting out only the pages you need using PDF Split to reduce size by removing pages rather than by compressing.
The PDF needs to be split first
Sometimes the fastest way to reduce a PDF's size is to remove pages you do not need. A 40-page bank statement compressed to 60 percent of its size is still 24 pages of content. If the recipient only needs pages 8 and 9, extract those with PDF Split first, then compress if needed. Both tools run in your browser with no upload.
Other PDF tools that run in your browser
SammaPix offers a full suite of browser-based PDF tools, all with no upload and no server processing. Here is when to use each:
- PDF Compress: reduce file size by rasterizing and re-encoding page images. Best for scanned documents and image-heavy PDFs.
- PDF Split: extract specific pages, split into individual pages, or divide into chunks. Covered in detail in How to split a PDF privately, no upload.
- PDF Merge: combine multiple PDFs into one. Covered in How to merge PDFs privately, no upload.
- JPG to PDF: combine one or more images into a PDF. Useful for creating a document from scanned photos.
All your PDF needs, all in-browser
Compress, split, merge, and convert PDFs without uploading them anywhere. All tools run locally in your browser. No server. No signup. No watermark.
FAQ
Does compressing a PDF online mean uploading it to a server?
With most tools, yes. iLovePDF, Smallpdf, Adobe Acrobat Online, and similar services upload your PDF to their servers for processing. With SammaPix, no. The compression runs entirely in your browser using pdf.js (to render pages) and pdf-lib (to build the output PDF). Your file never leaves your device. You can verify this yourself by opening your browser's Network inspector (F12) and watching for outgoing requests while the tool processes your file. You will see none.
Why does browser-based PDF compression work better on scanned PDFs?
Browser-based PDF compression works by rasterizing each page (rendering it as an image) and re-encoding it as a JPEG at a lower quality. Scanned PDFs are already image-based: each page is a high-resolution scan embedded in the PDF container. Re-encoding that image at a lower JPEG quality produces a noticeably smaller file. For example, a 15 MB scanned invoice bundle compressed at Medium quality can easily become 4 to 6 MB, which is a 60 to 73 percent reduction. Text-only PDFs, by contrast, store text as vector data (actual characters with font information). Rasterizing vector text and re-encoding it as JPEG does not produce much size reduction, and can make the file larger than the original because JPEG encoding of sharp text is not efficient.
Will the compressed PDF still have selectable text?
Not if the PDF is compressed using the rasterization method. When each page is rendered as an image and re-encoded as JPEG, the output PDF contains images of the pages, not the underlying text data. Text in the compressed PDF will not be selectable or searchable. This is an honest trade-off: you gain significant size reduction on image-heavy or scanned PDFs, but lose text selectability. If you need selectable text in a smaller PDF, the right tool is a server-side compressor that uses font subsetting and stream compression, such as Ghostscript. The SammaPix tool shows you the before and after file size so you can decide whether the trade-off is worth it for your use case.
What PDF types compress best with this tool?
The tool works best on: scanned documents (the most common case for large PDFs), PDF portfolios with embedded photos, brochures and catalogues with full-page photographs, presentations exported to PDF with image-heavy slides, and WhatsApp or iMessage PDF attachments that are actually photos. The tool produces limited reduction on: text-only PDFs (legal documents, ebooks, plain reports), PDFs that already use aggressive JPEG compression internally, and PDFs with mostly vector graphics.
What compression quality level should I choose?
Medium is the recommended starting point for most documents. It typically reduces a scanned PDF by 50 to 70 percent while keeping images readable. Low produces the smallest file but images may look blurry or blocky at close range — acceptable for archiving, not ideal for sharing with clients. High preserves near-original image quality with a smaller file reduction (usually 20 to 40 percent). For email attachments, Medium or Low is usually sufficient since recipients rarely zoom in to full resolution.
Is there a file size limit for the PDF Compress tool?
The practical limit is your device's available memory, since the entire PDF is processed in the browser. For most PDFs under 100 MB the tool works fine on any modern device. Very large PDFs (several hundred MB) may be slow or may hit browser memory limits on older or low-RAM devices. There is no artificial file size cap imposed by the tool itself, because no server is involved.
How do I verify no upload happens?
Open your browser developer tools (F12 on Windows/Linux, Command Option I on Mac), click the Network tab, then drop your PDF into the SammaPix PDF Compress tool and click Compress. Watch the network panel. You will see requests for static page assets (JavaScript, CSS) when the tool loads. During compression, you will see zero outgoing requests. The PDF is read by the FileReader API, processed entirely in memory, and the output is downloaded via a blob URL — no network call is made.
