Drop in your PDFs
Several files at once is fine. Parsing, re-encoding and writing all happen in this browser.
ImgIng / PDF compression
SELF-BUILT PDF ENGINECompress PDFs with two-axis effective DPI and content-aware cleanup
ImgIng never round-trips a PDF through HTML or flattened page images. It optimizes the original structure locally, makes per-image decisions and reclaims resources that actual drawing operations do not use while preserving page geometry, text, links and vectors.
No PDF uploadTwo-axis effective DPIContent-aware cleanupOutput verified by re-reading
Capability quick facts
These facts describe the current product, not unshipped roadmap work.
- Where it runs
- In your browser; no HTML reconstruction
- Decision basis
- Two-axis effective DPI plus actual resource use
- Structural work
- Inline images, unused resources, stream merge and Flate race
- Stated limits
- No font subsetting; encrypted PDFs must be unlocked first
Reviewed by the DataDance product and engineering team · Published · Updated
Finish in three steps
Confirm important settings before export; the processing location is always disclosed.
Pick a preset
Screen 72dpi, e-book 150dpi, print 300dpi, or lossless structure only. You can also name a target file size.
Compare and save
Switch between the original and the result in the page, then save to your device once it looks right.
Why is PDF compression still mostly about images?
A PDF contains text, vectors, fonts and bitmaps. Scans, photos and screenshots usually dominate its size. ImgIng finds ordinary Image XObjects and promotes safely parsed BI / ID / EI inline images into the same pipeline before deciding what each image deserves.
Why measure effective DPI independently on both axes?
The same image can be reused at different sizes or placed with non-uniform scaling. ImgIng follows page and Form graphics matrices, aggregates the maximum horizontal and vertical placement independently, and chooses each target dimension separately. Component Decode arrays are applied before re-encoding, while colour spaces that require full colour management remain untouched when equivalence cannot be proven.
Why not turn every image into JPEG?
JPEG suits photographs but can be larger and blurrier for screenshots, charts and flat colour. ImgIng races JPEG, indexed-palette Flate, grayscale Flate, RGB Flate and the original. Palette dictionary overhead is included in the comparison, and only a genuinely smaller candidate wins.
What happens beyond images?
Actual drawing operators drive conservative removal of unused fonts, images, graphics states, colour spaces, patterns, shadings and properties. Unknown operators or special streams keep their complete resource sets. Fragmented page content streams merge only across safe lexical boundaries and only when the result is smaller. Uncompressed, ASCII, LZW and existing Flate streams all race a fresh Flate encoding; unreachable-object collection, exact stream deduplication, metadata cleanup and object-stream packing then run as before.
What will ImgIng refuse to touch?
Encrypted or permission-protected PDFs must be unlocked first. JPEG 2000, JBIG2, CCITT fax, CMYK print JPEG and images needing full ICC, Lab or DeviceN colour management remain original when the browser cannot re-encode them equivalently. Any image, structural stream or complete PDF that does not become smaller also keeps its original data.
Honest limits: ImgIng does not currently subset fonts, so text-only gains depend on unused resources and inefficient structural streams. The output is re-read to verify page count, every page box, selectable-text operators and link semantics; any mismatch keeps the source PDF. Rewriting a digitally signed PDF invalidates the signature.
Updated 2026-08-25 · Live capability detection inside the tool is authoritative
Frequently asked questions
These visible answers match the current product behaviour and structured data.
Does my PDF get uploaded?
No. Parsing, image re-encoding, structural rewriting and verification all happen in this browser, and no extra runtime needs to be downloaded.
Is the text still searchable and selectable afterwards?
Yes. ImgIng only re-encodes embedded bitmaps and rewrites the file structure. Pages are never flattened into images, so the text layer, bookmarks and links survive.
Why do some PDFs barely shrink?
If the embedded images are already below the chosen DPI and sensibly encoded, or the file is mostly text and vectors, there is nothing to reclaim. ImgIng says so and keeps your original rather than manufacturing a number by lowering resolution.
Explore ImgIng
Each task page documents real settings, limits and format advice—not keyword-swapped duplicates.