About this no-upload PDF compressor
LocalPdfCompress is a small, English-only utility site: Compress PDF online free — no upload. You drop a PDF that is already on the device, choose a structural rewrite mode, then download a new file when you want to compare sizes. Compress runs in your browser with pdf-lib. There is no account, no document library, no remote “paste a URL” box, and no advertising on this pages.dev experiment.
Who publishes this site
LocalPdfCompress is a static website on Cloudflare Pages at localpdfcompress.pages.dev. It is not a cloud editor, not a signing product, and not a document drive. The public pages you can open today — the tool, four guides, this about page, privacy, terms, contact, and the trust check — are the whole product surface. English is the only language. There is no custom domain in front of this experiment, and no cross-promotion of other products on these pages.
Corrections and accessibility reports go to contact@localpdfcompress.pages.dev. That mailbox is for wording, broken controls, and honest-limit mistakes. It is not a private converter and not a place to send contracts or ID scans. See contact.
What the PDF compressor is for
The jobs are ordinary and specific. People search for compress PDF, PDF compressor, or reduce PDF file size when they want a smaller attachment — not when they need OCR, e-sign, or certified redaction. Typical cases:
- Trying a local rewrite before sending a PDF that feels larger than it needs to be.
- Comparing before/after sizes honestly, and keeping the original when the rewrite grew.
- Avoiding an upload-based compressor for a draft that should stay on the device during the process step.
- Understanding that structural rewrite is not the same as image downsampling.
The home page is the tool. Four guides walk those cases without inventing new URLs:
- 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
What this version does not do
PDF files only. Word, images, and remote URLs are rejected or out of scope. We do not market OCR, merge/split-as-hero, password cracking, DRM bypass, certified redaction, e-discovery, or regulatory compliance (including any 等保 claim). There is no server fallback when the tab runs out of memory.
Soft guidance is about 25 MiB per file, because the work uses your browser’s memory. A phone can fail sooner. The output is a new page-copy document: digital signatures, bookmarks, forms, and accessibility tags are not guaranteed to survive. Embedded images are not downsampled in this version. Those numbers and limits are honesty, not marketing.
How it is built (and how you can verify)
PDFs are read with the browser File API. Pages are copied with pdf-lib (MIT) running in the page. The script is served from this host as /assets/vendor/pdf-lib.min.js, so the page does not need a third-party script host for the engine. Your document bytes are not part of that request. Downloads use a blob: URL created in the tab.
The privacy policy says what static hosting can still log. The trust page walks through a DevTools → Network check so you can verify that compress does not POST the file. The terms cover rights in your documents and the limits of a free browser tool.
Hosting is Cloudflare Pages. Optional Cloudflare Web Analytics may record aggregate page views. It is not a document store.
Visual identity
This site uses a teal + coral palette (primary #0f766e, accent #ea580c, background #f0fdfa) with a circular compress-gauge hero — intentionally different from sibling PDF tools that use terracotta/olive stacks, rose cards, or slate+lime thumbs.
Content and advertising stance
This experiment does not run advertising scripts today. The HTML we ship does not include an ad network. Response headers set a Content-Security-Policy with script-src 'self' that does not list an ad host. If that stance ever changes, the privacy policy and this page would need an honest update first. Until then, treat “no ads on this pages.dev experiment” as the published rule.
Compress for email, ZIP, and long-tail intent
People who land here from “compress PDF for email” or “reduce PDF file size without upload” still get the same local tool. There is no second product and no typed attachment guarantee. Structural rewrite may help packing; it will not downsample images. ZIP is a different job (transport wrapping), covered on the home page so we do not invent a fake archive feature.
Honest engine bounds live in limits. Privacy during the process step lives in why local compress matters. Soft ~25 MiB guidance, no password bypass, and no claim that every file shrinks are repeated on purpose — AdSense-style depth should not become capability fiction.
Corrections
If a limit, a privacy sentence, a guide example, or a control label is wrong, write to contact@localpdfcompress.pages.dev. Quote the page URL and the sentence. Please do not attach contracts, ID scans, medical letters, or unredacted packets. We cannot recover a file that never left your device. A clear description of what you expected is enough.