Foxit offers two built-in paths to compress a PDF — a one-click Reduce File Size command and a deeper PDF Optimizer Advanced mode — and both run inside the Foxit desktop app without uploading your document. The Reduce File Size command targets quick shrinks by downsampling raster images and re-encoding compatible streams, while PDF Optimizer Advanced exposes granular controls for color images, grayscale images, fonts, transparency flattening, and object cleanup. Both modes preserve the page count, text layer, and form structure, but they do rewrite image byte streams at quality levels Foxit picks, and they cannot reach every embedded object type or every color space the file contains. For users who want a smaller file and either do not have Foxit installed or want a result that never leaves the browser tab, a local browser tool can re-encode eligible JPG images inside the current page and verify the output before any download is offered to the reader.
What Foxit's Reduce File Size Command Does
Foxit PDF Editor exposes two compression paths in its desktop app, and the older one is the one-click Reduce File Size command. For versions before V10, the menu path is File → Reduce File Size. For V10 and later, the same command moved to File → PDF Optimizer → Reduce File Size, and Foxit also added a Multiple Files submenu that lets you add several documents, reorder them, set an output folder, choose a file-naming pattern, and decide whether to keep the current PDF version or downconvert to an older one.
The Reduce File Size command is the fastest option. It rewrites compatible image streams at quality levels Foxit chooses, downsamples high-resolution rasters, removes redundant objects, and rewrites the cross-reference table. Text, links, form fields, and the page count stay intact. Before committing, the dialog shows an estimated compressed size so you can judge whether the pass is worth the quality trade, and you can choose to retain the current compression settings or switch to a specific PDF version.
PDF Optimizer Advanced: The Deeper Pass
For finer control, the same File menu offers PDF Optimizer Advanced. This mode opens a settings panel with separate quality sliders for color images, grayscale images, and monochrome images, plus independent controls for fonts, transparency flattening, user data, bookmarks, and metadata. You can target a specific PDF version for output and choose to discard form data, comments, embedded thumbnails, or document JavaScript actions before saving.
The advanced pass takes longer because it inspects more object types, but it reaches objects that the one-click command skips. If your PDF came from a scanner, an OCR layer, or a design tool that produced rich image objects, the advanced mode is more likely to find savings. It is also the right choice when you need to enforce a target PDF version for archival or regulatory workflows, because the dialog lets you lock the output to a specific version instead of leaving it implicit.
Compress a PDF Locally in Your Browser
If you do not have Foxit installed, or you want a compression pass that never leaves your browser, the Compress PDF tool runs entirely inside the current tab. Reading, image decoding, re-encoding, output assembly, and verification all happen on your machine, so the source PDF is never sent to a remote server.
- Open the Compress PDF page and choose one local PDF file up to 25 MB from your device.
- Select Compress PDF and wait while the tool enumerates embedded image objects, decodes eligible JPG streams, and re-encodes them locally at JPEG quality 0.70 with the longest side capped at 1400 pixels.
- Read the reported input and output sizes. If the output is smaller and the verification pass succeeds, the Download button activates and you save the new PDF.
The tool only changes unmasked DCTDecode JPG images with DeviceRGB or DeviceGray color data. Everything else — text, fonts, vectors, forms, annotations, metadata, and the page count — stays exactly as the source had it. If no eligible image becomes smaller, the tool returns an honest error and does not produce a download, so you never see a misleading "compressed" file. If you want to compare how this local approach differs from a generic file-size compressor, the guide to compressing a PDF locally without uploading walks through the same workflow in more depth.
Which Files Actually Get Smaller
The browser tool targets a narrow slice of PDF content, and understanding that slice explains why some files shrink a lot while others barely move. Use the table below as a directional guide; the exact percentage depends on the source image bytes and PDF overhead, and the tool always reports the real input and output numbers after each run.
| PDF Type | What the Tool Can Change | Expected Direction |
|---|---|---|
| Photo album or scanned pages with raw JPG images | Eligible DCTDecode JPG with RGB or Gray color | Often substantially smaller |
| Marketing flyer with mixed JPGs and PNGs | Only the JPG objects; PNGs untouched | Moderate shrink |
| Text-only report, invoice, or e-book | No eligible image objects | No shrink; tool reports error |
| PDF with CMYK images, JPX streams, or masked art | Skipped by design | Usually no shrink |
| PDF with 300 DPI screenshots or higher | Downscaled to 1400 px longest side | Often noticeably smaller |
The shape of the table matters more than any single number: the tool rewards files that are heavy in raw JPG bytes and punishes files that lean on text, vectors, or unusual color spaces. If your file falls in the lower rows, you need a general-purpose optimizer rather than a JPG re-encoder.
Safety Limits and Verification Inside the Browser
The local tool imposes three hard caps to keep decoding predictable and to prevent runaway memory use. It accepts one PDF up to 25 MB, processes at most 80 eligible image objects, and stops at 100 megapixels across the images it decodes. Files outside those limits return a clear error before any work begins, which is why the tool is bounded rather than open-ended.
Every replacement is gated by a byte comparison. The tool decodes each eligible JPG in a browser image element, downsamples its longest side to 1400 pixels, re-encodes at JPEG quality 0.70, and only swaps the stream in the PDF when the candidate is genuinely smaller than the original. The new payload replaces the original at the same indirect reference, so page layout, page count, and references to the image stay untouched. The re-encode uses the browser's canvas pipeline, which follows the HTML Canvas toBlob encoding path described in the WHATWG HTML standard.
Before the Download link appears, the tool reopens the assembled PDF with the PDF.js library and confirms the page count matches the source. The Mozilla PDF.js API reference documents how the library loads and parses a PDF inside the browser. A malformed output is rejected before you can save it, so a broken PDF never reaches your downloads folder.
When to Stay With Foxit Instead
Foxit's desktop app still has advantages the local browser tool does not match. It handles CMYK image streams, JPX-encoded objects, transparency groups, font subsetting, and metadata cleanup. It can target a specific PDF version, strip user data, and remove bookmarks. It also processes files larger than 25 MB, which the bounded browser tool refuses. If you need archival conformance, print color management, or guaranteed lossless behavior, use Foxit's PDF Optimizer Advanced mode in the desktop app or another specialist PDF optimizer.
For routine photo-heavy documents — scans, slide decks exported as PDFs, product catalogs, image-heavy reports — the local Compress PDF tool is a faster path because there is no install, no upload, and no waiting on a cloud queue. Keep the original file, inspect the output in a PDF viewer, and confirm the visual quality is acceptable before sharing it. If you also need to watermark the result the same way you would inside Foxit, the local watermark workflow for Foxit users covers that workflow without uploading either.