Reduce PDF file size for upload limits (LinkedIn, IRS, USCIS)

Upload limits at the services people actually hit

Limits taken from each service's public help documentation as of May 2026 and verified by test-upload. Subject to change without notice.

Step-by-step: shrink a PDF to fit a specific cap

The ScoutMyTool tool lives at scoutmytool.com/pdf/compress-pdf. Runs client-side via pdf-lib — no upload, no signup, no quota.

1. Drop your PDF. One file at a time. The file loads into a sandboxed memory buffer; nothing is uploaded.
2. Pick a preset, OR set a target size. Three presets (Light, Medium, Aggressive) cover most cases. For "compress until under N MB", use the "Target size" field — the tool iterates downsampling and re-encoding until the output is under the target or the floor is reached.
3. Click Compress. The tool walks every embedded image, downsamples to the preset's target DPI, re-encodes to JPEG at the preset's quality, then re-stamps the cross-reference table and font subset cache. Progress shown live; typical 20-page document with 5 embedded photos compresses in 5–15 seconds.
4. Review the size delta. The result panel shows before / after sizes side-by-side with the percentage reduction. If the result still exceeds your target, switch to a more aggressive preset or set an explicit target size.
5. Verify legibility. Open the output and zoom in to the smallest text. If it is still sharp at 100% zoom, the compression was safe. If text is blurry, back off the preset.
6. If a target cap still cannot be hit. The file likely contains very large embedded images that dominate the size. Open in PDF Editor and remove non-essential ones, or split via Split PDF and submit as multiple attachments (works for IRS, USCIS, Gmail).
7. If the source is password-protected. Unlock first via Unlock PDF.

What "compress PDF" actually does

PDF file size is dominated by three things: embedded images, embedded fonts, and the binary stream wrappers that wrap everything. The PDF specification (ISO 32000-1) defines several compression filters that can be applied to streams — FlateDecode (deflate / gzip-style), DCTDecode (JPEG), JBIG2Decode (bitonal documents), JPX (JPEG 2000), etc.¹.

• Images typically account for 70–90% of file size in any PDF that contains photographs. Downsampling (reducing the pixel count) and re-encoding (JPEG at lower quality) are the biggest levers.
• Fonts account for 5–20% in form-heavy PDFs. Subsetting (keeping only the glyphs actually used) is the standard optimisation; the tool re-subsets and merges duplicate-font tables.
• Structure overhead (cross-reference table, page tree, object boundaries) accounts for the rest. Re-stamping the cross-reference table can save 1–3% on its own and is always safe.

U.S. government accessibility guidance (Section 508) recommends 150 dpi as the minimum resolution for embedded images in submitted PDFs² — which is the floor the Medium preset uses by default and the reason it works as a single-button choice for most filing workflows.

Related PDF tools on ScoutMyTool

• Compress PDF — the tool this guide is about.
• Compress PDF guide — broader compression walkthrough not specific to upload limits.
• Split PDF — break the file at natural boundaries when compression alone is not enough.
• PDF Editor — remove specific bulky images / pages.
• Send Large PDF Email — workflow for files that need to clear email caps.
• PDF email size limit guide — per-provider cap reference.
• Unlock PDF — required first if the source is password-protected.

Frequently asked questions

My PDF is 12 MB and LinkedIn caps resume uploads at 2 MB. What do I do?

Pick the "Aggressive" preset, click Compress, download. A typical 12 MB resume PDF that was exported from Word with full-resolution embedded photos drops to about 1.5–2 MB on the Aggressive preset, well under the 2 MB cap. If the result is still over 2 MB, the file is dominated by something other than image resolution — usually embedded fonts or a hidden full-page background image. Switch to PDF Editor, remove the background image, re-export. If the resume is mostly photo content (designer portfolios), drop the photo resolution to 96 dpi specifically rather than relying on the global preset.

The IRS rejected my attachment for being too big. What is the actual limit?

For most IRS e-file attachments (Form 8453 supplemental materials, Form 1040 cross-references), the practical per-file limit is 3 MB, with a 60 MB total per return. Use the "Medium" preset first — that hits 3 MB on most W-2 and 1099 stacks without visibly degrading text legibility. If a single document exceeds 60 MB after compression, split it via Split PDF and submit as multiple attachments. The IRS publication accepted file-format guidance is at the reference at the bottom of this article.

USCIS limits each upload to 6 MB. My I-130 packet is 18 MB.

Two passes: (1) Compress PDF with the Medium preset usually brings a typical I-130 packet from 18 MB to 4–5 MB by dropping embedded passport photos and supporting-evidence scans from 300 dpi to 150 dpi (still well above USCIS's minimum legibility requirement of 100 dpi). (2) If still over 6 MB, split via Split PDF into the natural document boundaries (the petition itself, the supporting evidence per document type) and submit them as separate uploads — USCIS's case-management portal accepts multiple attachments per filing.

Which preset should I pick — Light, Medium, or Aggressive?

Light keeps embedded images at near-original quality and only re-optimises the PDF structure; expect 10–20% size reduction; right for "PDF is slightly over the limit, do not want any quality loss". Medium downsamples images to 150 dpi and applies moderate JPEG re-compression; expect 40–60% reduction; right for most upload-limit fixes — the visual difference is minimal on screen and acceptable in print. Aggressive downsamples to 96 dpi (screen resolution) and applies harder compression; expect 70–85% reduction; right when the document will only ever be viewed on screen and file size is critical (LinkedIn 2 MB, Twitter 5 MB).

Is my PDF uploaded to your servers?

No. Compression runs entirely in your browser using pdf-lib and the canvas-based image re-encoder. Your file is loaded into a sandboxed memory buffer, embedded images are downsampled and re-encoded locally, the modified file is delivered as a download. Verify in DevTools Network — no outbound requests during compression. This matters when the PDF you are trying to upload-to-LinkedIn-or-USCIS contains personal data (driver licence scans, passport pages, tax documents) that should not be uploaded to a third-party "compressor" first.

Will compression change the page count or layout?

No. Compression operates on image streams and font streams within each page — it does not add, remove, or reorder pages, and it does not reflow text. The compressed PDF has the same page count, the same page dimensions, the same text positions, and the same form-field layout as the input. Only the binary representation of embedded raster images and fonts changes. Open the input and output side-by-side; the difference is in file size, not in what is on each page.

Can I make the PDF smaller without any visible quality loss at all?

Sometimes. Pick the "Light" preset and toggle "Lossless only". This re-encodes image streams using Flate (deflate) compression rather than JPEG, removes duplicate font subsets, and re-stamps the cross-reference table — all without any pixel-level changes. Typical savings: 5–20% on text-heavy PDFs, less on photo-heavy. If you need lossless AND a large reduction, the PDF probably contains content that genuinely cannot be losslessly compressed further (already-compressed photos at the JPEG limit). At that point, either accept some quality loss via Medium or restructure the document upstream.