Optimize & Repair PDF
Compress PDF files with Extreme, Recommended, or Low settings. Repair corrupted PDFs, flatten form fields and annotations, and edit document metadata — all processed client-side.
Optimize & Repair PDF
Compress, repair, flatten, or edit metadata of a PDF — all processed client-side in your browser.
Drop a PDF file here
or click to browse
How to Use the PDF Optimize and Repair Tool
A project manager uploads a 45 MB PDF proposal to the company file server but the upload times out because the attachment limit is 25 MB. Open this tool, drag the PDF onto the drop zone, select the Recommended compression level, and click Compress and Download. The tool strips non-essential metadata, removes unused objects, and applies zlib stream compression to the internal data streams. The result downloads as a smaller file that fits within the email attachment limit without changing any visible content.
A user reports that a PDF from a client opens with a parser error in Adobe Acrobat. Upload the file, switch to the Repair tab, and click Repair and Download. The tool loads the file with error-tolerant parsing, rebuilds the cross-reference table from scratch, re-indexes every object, and writes a fresh standards-compliant PDF. The repaired file opens cleanly in all readers.
A signed PDF contract contains interactive form fields that some recipients cannot view correctly in their PDF reader. Upload the file, switch to the Flatten tab, and click Flatten and Download. The tool merges all form fields, annotations, and interactive elements into the page content layer, producing a flat non-editable PDF that renders identically in every viewer.
A whitepaper PDF still has the author field set to the intern who drafted it three years ago. Upload the file, switch to the Metadata tab, update the Author and Title fields, add relevant keywords, and click Save Metadata. The tool reads the existing metadata, applies your changes, and downloads a new PDF with the corrected information embedded in the document properties.
A university library needs to store 10,000 scanned research papers on a limited-budget server. Upload a representative sample, select the Extreme compression level, and click Compress and Download. The tool strips all metadata including Title, Author, Producer, and Creator fields, removes non-essential objects, and applies maximum stream compression. Text-heavy PDFs typically shrink 10 to 20 percent, while image-heavy documents can see 30 to 50 percent reduction.
Related tools
How the PDF Optimize and Repair Tool Works
The tool rebuilds PDF internals using pdf-lib to shrink file size or restore structural integrity. All processing happens in your browser with no server calls.
Compression Techniques
During compression the PDF is re-serialised with object streams, which deduplicate shared resources and apply zlib stream compression. Image-heavy documents benefit from recompression, while text-heavy files gain from font subsetting and removal of unused objects. The Extreme level additionally strips all metadata including Title, Author, Producer, and Creator fields, and discards non-essential entries. The Low level preserves all metadata and applies only structural optimisation for minimal size reduction.
Repair Algorithm
The repair mode loads the file with error-tolerant parsing using the ignoreEncryption flag, then writes a fresh PDF from scratch. This rebuilds the cross-reference table, re-indexes every object, and corrects malformed streams. Broken links between pages and their content streams are resolved, producing a valid standards-compliant document. The repair process does not modify visual content, it only fixes the internal document structure.
Flatten Process
Flattening iterates through every page and removes Widget annotations, which represent form fields, from the page annotation arrays. The AcroForm dictionary is deleted from the document catalog, removing the form field definitions entirely. The visual appearance of the form fields is preserved as part of the page content layer, producing a flat non-editable PDF that renders identically in all viewers.
Frequently Asked Questions
How much can file size be reduced?
Reduction varies by content type. Text-heavy PDFs typically shrink 10 to 20 percent through structural compression and object deduplication. Image-heavy documents can see 30 to 50 percent or more at the Extreme level because the tool removes metadata and non-essential objects. The actual savings depend on how much redundancy exists in the original file.
Will compression decrease image quality?
The tool does not re-encode embedded images. Compression comes from deduplication of shared resources, stream-level zlib compression, and metadata removal. Visual quality remains unchanged at all compression levels because the image data streams are preserved exactly as they are in the original document.
Can it fix corrupted PDFs?
Yes. The repair mode rebuilds broken cross-reference tables, re-indexes objects, and corrects malformed streams by writing a fresh PDF from scratch. Severely damaged files with missing page content or truncated streams may not be fully recoverable, but most common corruption issues like broken cross-reference tables and malformed objects are fixed.
Are bookmarks preserved during compression?
Bookmarks and the document outline are preserved at the Recommended and Low compression levels because these levels retain the document structure. The Extreme level strips all metadata and non-essential objects, which may remove bookmarks along with other metadata entries. If you need to preserve bookmarks, use the Recommended level instead.
Does it support encrypted PDFs?
The repair mode can load password-protected files using error-tolerant parsing with the ignoreEncryption flag. Compression and flatten require the file to be unencrypted, so remove the password first if possible. You can use the Secure PDF tool to decrypt the file before running optimisation.
Is my PDF uploaded to a server?
No. Every operation runs entirely in your browser using pdf-lib. Your file never leaves your device, ensuring complete privacy. All compression, repair, flatten, and metadata operations happen in browser memory with no network calls whatsoever.