LocalPdfCompress

Compress PDF online free — no upload, no sign-up

This page is a browser PDF compressor for a file already on the device. Search intent is compress PDF / PDF compressor with a privacy-first local path — not merge, not image-to-PDF, not PDF-to-image. There is no upload step and no account. The tab reads the file, rewrites pages with pdf-lib, and lets you download a new PDF when you want to compare sizes.

1. You pick one PDF with the file control, or you drop it. The File API reads that local file. Loading the page fetches HTML, CSS, and the pdf-lib script from this host. That is the app, not your document.
2. Choose Structural rewrite or Structural rewrite + strip Info metadata. Both copy pages into a new document and save with object streams. Metadata strip clears common Info fields and sets a short producer string — it does not redact page content.
3. Press Compress on this device. The download link is a blob: URL created in the page — not a file fetched from our server.
4. Read the before/after sizes. If the rewrite grew, keep the original. Image-heavy scans often need image downsampling, which this version does not do.

Open the download before you attach it. Check content and page count. Digital signatures, bookmarks, forms, and accessibility tags are not guaranteed to survive a page-copy rewrite. Prefer a laptop for large files; phones hit memory limits sooner.

This version will not fetch a PDF from a URL, keep a cloud copy, run OCR, merge or split packets, convert images, crack passwords, or bypass DRM. Use files you have the rights to process. How to prove the document stayed in the tab: verify PDFs stay on this device.

Compress a PDF for email without an upload converter

A common job is one smaller attachment: landlord packet, clinic form, teammate handoff, or a ticket appendix. Upload converters receive a full copy of the file before they shrink it. A local rewrite builds a candidate attachment in the tab, then you choose whether the after size is worth sending.

There is still no typed “under 2 MB for Gmail” guarantee. Soft guidance is about 25 MiB for the input. If the after size is still too large for your mail host, that is a product limit of structural rewrite — not a server queue you can raise. Image-heavy scans usually need a tool that downsamples images; this MVP does not. Steps without inventing a new product: compress without uploading and what actually changes when you reduce size.

Compress PDF vs ZIP — different jobs

Zipping a PDF wraps the same bytes in an archive. Recipients still need to unzip, and many mail clients treat a ZIP of a PDF as less convenient than a PDF attachment. Structural rewrite changes how the PDF itself is packed (object streams, page copy, optional Info strip). It does not replace ZIP for bundling many files, and ZIP does not replace a compressor when the recipient expects a single .pdf.

If you only need transport wrapping, use your OS archive tool. If you need a PDF that opens without unzipping, try the rewrite here and keep the smaller of before/after. Neither approach downsamples embedded images by itself. Honest engine bounds: PDF compressor limits.

What “compress” means here (honest)

Many online “PDF compressors” re-encode images or apply server-side tools. LocalPdfCompress is different on purpose:

• Implemented: browser-side rewrite with pdf-lib — page copy, object streams, optional Info metadata strip. Before/after size is shown so you can decide.
• Not implemented: embedded image downsampling, JPEG recompress of page bitmaps, ghostscript-style binary recompress, or a slider that promises “email size under X MB.”
• Why say so: pdf-lib is excellent at creating and copying PDF structure. It does not replace a full raster pipeline. Claiming “always half the size” would be false.

Guides that expand the same facts without inventing features:

• Compress a PDF online without uploading
• Reduce PDF file size — what actually changes
• PDF compressor limits (structural rewrite)
• Why a local PDF compressor matters for privacy

Honest limits for no-upload PDF compress

• Formats: PDF only (.pdf). Word, images, and remote URLs are out of scope for this tool.
• Memory, not a quota: soft cap about 25 MiB per file. A phone can still fail under the cap. Decoding a large PDF takes more RAM than the file size on disk.
• Page copy: the output is a new document built by copying pages. Digital signatures, bookmarks, forms, and accessibility tags are not guaranteed to survive.
• Encrypted PDFs: password-protected files may be rejected. This is not a DRM or password bypass tool.
• Not in this version: image downsampling, OCR, merge/split-as-hero, e-sign, accounts, and ads.
• Sensitive documents: local processing reduces exposure versus upload converters, but a browser extension or other software on the device can still see tab memory. See privacy guide and trust.

Publisher context, contact mailbox, and policy pages: about · contact · privacy · terms.

Worked example: compare before you keep

You have packet.pdf at 4.2 MiB. Drop it, choose structural rewrite, press Compress on this device. If the after size is 3.9 MiB, download and open it — skim page count and a few pages. If the after size is 4.4 MiB, discard the rewrite and keep the original. That outcome is normal when the PDF was already lean or image-dominated.

Need merge, split, PDF↔image instead? Those are different products. This site only compresses by local rewrite. If Network ever shows a POST of your PDF after a drop, stop and write to the contact address with the page URL — without attaching the document. The normal check is on trust.

Why this pages.dev PDF compressor exists

LocalPdfCompress is a small English-only utility on Cloudflare Pages. It is not a document drive, not a signing product, and not a place to paste a remote PDF URL. The public surface is this tool, four guides, about, privacy, terms, contact, and trust — no new languages and no custom domain in front of the experiment today.

If a limit, a privacy sentence, or a guide step is wrong, email contact@localpdfcompress.pages.dev with the page URL and the sentence. Do not attach the PDF. We cannot recover a file that never left your device. More on who publishes the site: about.