100% Client-Side Zero File Uploads No Sign-Up Required

Client-Side PDF Compress, Merge & Split

Combine documents, extract specific page ranges, strip XMP/Info metadata, and compact PDF object streams locally — private enough for contracts, taxes, and IDs.

1. Load & Order PDFs

Click or drop PDF files here
Processed 100% locally with pdf-lib (MIT)
Add PDF files or click "Load Sample PDFs" to test Merge, Split, Inspect, and Optimize.

2. Inspect, Merge, Split & Optimize

Document Inspector
Select a PDF to inspect page count, dimensions, and metadata.

Merge Queued PDFs

Combines queued files in order using optional per-file page ranges.

Split / Extract Pages (Primary PDF)

Optimize & Clean Metadata (Primary PDF)

Phase 1 Android · Hybrid (AdMob + Pro Sub)

Processing >50 MB scans or need CameraX edge-detection scanning?

Install DocVault on Android for hardware-accelerated CameraX document scanning, bitmap downsampling to exact KB targets, and offline PDF vault locking. (net.ngoisaomedia.pdfvault)

How to Merge, Split, and Optimize PDFs Privately

  1. Drop one or more PDF files into the queue area—or click "Load Sample PDFs" to immediately generate a 3-page report and 2-page appendix locally for hands-on testing.
  2. Inspect any queued document to view its exact page count, page dimensions in points, byte size, and embedded Title, Author, Creator, and Producer tags.
  3. To combine documents, use the Up/Down buttons to arrange file order, optionally enter per-file page ranges (such as "1-2, 4"), and click "Merge Queued PDFs" to download a single combined file.
  4. To extract pages from the first PDF, type a range like "1-3, 5" and click "Split by Range" (or export every page individually). To reduce structural overhead and clean metadata, toggle Object Streams and Strip Metadata, then click "Optimize Primary PDF".

Structural Compaction, Range Validation & Transparent Metrics

Many web PDF tools impose strict two-file hourly limits, force account sign-ups, or upload confidential contracts to third-party cloud buckets. Our browser toolkit parses and rebuilds PDF object trees on your own CPU using pdf-lib. Merging copies only the requested page objects into a clean destination container, while splitting validates 1-based page ranges before generating fresh outputs.

During structural optimization, copying pages into a new PDFDocument automatically discards unreferenced orphaned objects left behind by incremental PDF editors, strips optional XMP/Info metadata dictionaries, and packs cross-reference tables into compressed object streams. We report exact before-and-after byte sizes honestly: if a PDF is already structurally compact (for example, dominated by pre-compressed high-DPI scans), we recommend DocVault on Android for raster image downsampling.

Zero-Upload Browser Document Processing

Financial statements, legal agreements, and identity scans should never traverse unknown cloud queues. Every byte read from your file picker stays inside an in-memory Uint8Array inside your browser sandbox.

Encrypted or password-protected PDFs are detected immediately with a clear message rather than failing silently, and no temporary copies are ever transmitted over the network.

Frequently Asked Questions

Why does structural optimization save more on some PDFs than others?

Structural compression removes orphaned incremental-save objects, strips XMP/Info metadata, and enables PDF 1.5 object streams without altering image quality. PDFs packed with large uncompressed bitmaps require raster recompression, which DocVault handles natively on Android.

Can I extract non-contiguous pages like 1-2 and 5?

Yes. Enter comma-separated page numbers and hyphenated ranges such as "1-2, 5, 7-9". Out-of-bounds or reversed ranges are caught immediately with clear feedback.

What happens if I upload a password-protected PDF?

The inspector cleanly reports that encrypted PDFs must be unlocked first, ensuring no silent corruption or partial output occurs.

Is there a daily file limit or file count cap?

No artificial limits are enforced by our site; performance is bounded only by your browser tab memory.