How PDF processing works

The process at a glance

All eleven PDF tools on this website follow the same basic pattern. This applies to compressing just as much as to deleting pages or converting images to a PDF. To do this, the website loads program code into your browser. Your documents are read and processed there, without the current version uploading them to a server for processing.

1. Select and check the file: You select a file on your device. The browser first checks the extension and size. If both are fine, it reads the file for a preview, depending on the tool.
2. Choose the task and settings: You decide what should happen and start processing. A result is only created from this moment on.
3. Check and save the result: You compare the original and the result and download the new file. It is only saved on your device once you download it.

In this process, your original file lies outside the processing path. The tool works with a copy in memory and does not overwrite anything on your device. You can find out which tool suits your task under Choose the right PDF tool.

Selecting and checking a file

As soon as you select a file, the browser first checks whether the name and size are acceptable. These quick checks run before it even reads the content:

• The file extension must match the tool, so .pdf for compression.
• The file must not be empty.
• A single PDF, JPG or PNG file may be at most 50 MiB, and a DOCX file for Word to PDF at most 20 MiB.
• If you select several files in a tool, together they may be at most 100 MiB.

Depending on the tool, the browser already creates a preview at this point and reads details such as the page count or the image dimensions. Selecting a file alone, however, does not yet compress, change or save anything.

An extension alone doesn't make a valid PDF

The .pdf extension only tells you what a file is called. It does not tell you whether its content is a readable and supported PDF. Whether the tool can actually process the document only becomes clear from the check during the actual processing. So if only the preview fails, processing does not necessarily fail as well. Conversely, a file with a working preview may be rejected later because it has a property the tool does not process.

Program components are downloaded, your documents stay in the browser

For processing in the browser to work, the website downloads program components to your device: JavaScript, WebAssembly for compression and PDF.js for the previews. For text recognition, a dedicated worker, its program core and the language data for German and English are added when needed, and for Word to PDF the Liberation Sans font. All of these components come from the website itself. This is a download in your direction, and these requests contain no content from your file. The current version does not send your documents to a server for processing. How the data flows in detail and what this means for security is explained on the page File processing in detail.

Choosing the task and settings

Each tool has its own interface because the tasks differ. When compressing, a PDF should become smaller. When editing pages, you deliberately change the content. When converting, a different format is created, and the calculators only deliver numbers. You choose the settings and start processing with the respective tool's button.

Only when compressing is a smaller file the goal. When merging, splitting, rotating or converting, the result can be larger, smaller or about the same size. What matters there is whether the pages and content are correct. Merging, splitting, deleting pages and rotating transfer the pages into a new PDF. In the process, bookmarks, annotations, metadata, links or a PDF/A or PDF/UA identification may change or be lost. What is preserved is shown by the notes in the tool and by the saved file. When exporting as JPG or PNG, by contrast, the tool renders the pages as images. These image files contain no PDF features such as links, bookmarks or form fields.

The three compression methods

The compressor offers three methods. They differ mainly in whether images are changed:

• Lossless optimization rewrites the internal structure of the PDF, meaning its objects and data streams. Images are not re-encoded in the process.
• Balanced and Smaller file also optimize the structure. In addition, they may re-encode suitable embedded images as JPEG. Not every image is suitable for this, and not every PDF contains such images.

None of the methods rasterizes whole pages or removes pages. None of them changes the resolution of scans or adds text recognition (OCR). How much a PDF shrinks therefore depends on what it is made of. A file with content that is already heavily compressed may stay almost the same size. How the methods affect quality is explained in the guide Compress PDF without losing quality. You can find step-by-step instructions directly at Compress PDF.

Setting a target size

With the Set target size option, you specify an upper limit in KB, MB or MiB. On the home page it is initially switched off; on the page Reduce PDF to a target size it is already switched on. The target counts as reached if the result has at most as many bytes as the limit.

Smaller PDF or ZIP archive?

If you put a PDF into a ZIP or 7z archive with an archiving program, you don't get a smaller PDF but a different file format. So if a portal or an office explicitly requires a PDF, an archive won't help you directly. The compressor always outputs a PDF. Only when splitting and when exporting as JPG or PNG may a ZIP be offered in addition, bundling the individual files. You can still save each of the individual files separately.

Calculating instead of editing

Five of the six calculators work only with numbers you enter yourself. The Check upload limits calculator can also read selected PDFs locally and takes over their exact file size. For PDFs up to and including 50 MiB, it also tries to read the page count. For larger files, it skips this step. If a PDF requires a password to open or cannot be read for another reason, the page count remains unknown. It still compares the size, and you can enter missing values by hand. In doing so, it does not change the PDF, does not measure upload speed and sends nothing to a recipient. You can find all calculators under Calculate file sizes.

Checking and saving the result

After processing, the tool shows file sizes, the page or file count and notes on the result, and many tools also show a preview. When compressing, you can compare the original and the result directly and move through the document with the page buttons.

Will my original file be changed?

No. The tool does not overwrite the file on your device. The result is a new file, and it is only on your device once you download it. Where you save it and whether you later replace the original with it is up to you. With a distinct file name, you can easily keep both versions apart.

How to check the saved file

1. Download the result and open the saved file: a PDF in a PDF program, exported JPG or PNG images in an image viewer, a DOCX file in a suitable word processor and a TXT file in a text editor. The preview in the browser alone is not enough.
2. Compare the page count with the original or with your selection.
3. Read the parts that matter: small print, tables, stamps, signature images and scanned pages.
4. Test links and form fields if your document contains any.
5. Compare the exact file size in bytes with the recipient's requirement.

When compressing, the tool also checks the result automatically: it reads the PDF structure and the pages, counts form fields and links, and detects whether structure information is present. Tools that create a new PDF reopen it as a check and verify that the page count is correct. For exported JPG or PNG images, there is no such PDF check. None of these checks proves that every feature works exactly as before, and no preview can guarantee that a recipient will accept the file.

Possible outcomes of compression

Every compression attempt starts from the original bytes. The tool keeps the smallest valid version. If it finds none that is smaller than the original, you get an unchanged copy. So you will never receive a file larger than before.

If the file stays unchanged or too large, the guide PDF won't get smaller will help you further.

At most or below?

The tool considers a target reached if the result does not exceed the limit. Some recipients, however, explicitly require a file below a limit. A hypothetical example: your set target size corresponds to exactly 2,000,000 bytes, and the result also has exactly 2,000,000 bytes. The tool reports the target as reached. A requirement of “smaller than 2,000,000 bytes”, however, is not met. In such cases, set the target size slightly lower.

Continuing to work with the result

Results do not move automatically from one tool to the next. If, for example, you want to delete pages first and then compress, you download the intermediate result. Then you select the saved file again in the next tool. There, that tool's own checks for format, size and content apply again: for a PDF the 50 MiB limit, and for a DOCX file in Word to PDF the 20 MiB limit. A DOCX file from PDF to Word is therefore not always accepted by Word to PDF: if it comes from a multi-page PDF, it contains several sections, and Word to PDF rejects such files.

Cancel, reset and try again

While processing is running, the tool shows the progress in processing steps. You cannot read a fixed duration from this, because it depends on the file and the device.

Cancel

The Cancel button appears while the tool is working. It stops the current processing and discards the unfinished result. Your file selection, the preview of your file and your settings are kept, so you can start a new attempt right away. When compressing, the background process is terminated, and when exporting as JPG or PNG, the tool stops rendering the page images. The other page tools stop at the next checkpoint between two processing steps. So it may take a moment until processing has actually come to a halt.

Reset

Reset goes further than Cancel, and the Edit new file button, which appears with the result after processing, does the same. Both cancel any running work, clear the selection, results, previews and download links, and reset all settings to their default values. If, on the other hand, you merely select a different file, for example via Change file, only the previous result and its download links disappear. The new file stays selected, and your settings remain as they were.

Why a download link no longer works

A result always belongs to exactly the file and settings it was created with. If you then change the file or an option, the previous result disappears and its download link becomes invalid. This prevents you from accidentally downloading a version that no longer matches your current settings. The same happens if you leave or reload the page: running work is cancelled, and the results are no longer available. So download a good result before you try something else.

Saved files are kept

Neither Cancel nor Reset deletes files you have already downloaded. They are stored on your device independently of the preview in the browser. The tool releases temporary working files after a result, an error or a cancellation. However, the website cannot determine when the browser actually frees up the memory used.

For another attempt with a different compression method, you don't need to reset anything. Simply change the setting and start again. Every attempt starts again from the original, so results are not compressed on top of each other multiple times.

Limits and error messages

Not every message has the same cause, so the same next step doesn't always help. A rejection concerns the file itself: it is too large, protected, signed or cannot be read as a PDF, and you need a suitable version. If, on the other hand, loading or processing in the browser fails, the document is not necessarily faulty; reloading the page or freeing up more memory is then more likely to help. If a PDF remains unchanged or above your target size, it was processed but could not be reduced enough.

Why the compressor rejects a PDF

• Encrypted or password-protected: The PDF is not broken but protected. The tools do not unlock passwords. You need an unprotected version, for example from the person or office that created the document.
• Digitally signed or certified: If the PDF contains a completed digital signature or a certification, the compressor rejects it. An inserted image of your signature or a scanned signature, by contrast, is ordinary page content and not a reason for rejection.
• Damaged, invalid or empty: The content cannot be read as a PDF. Try opening the file in a PDF program or get a fresh copy from the original source.
• Loading or processing error in the browser: If program components could not be loaded, reloading the page may help.

A digital signature is meant to show that a document has not been changed since it was signed. How this works is explained in Adobe's overview of digital signatures. So don't simply remove a signature just to get a tool to accept the file. The recipient might then no longer be able to verify it. The options available for signed documents and forms are described in the guide Compressing PDFs with signatures and forms.

Additional limits of the page tools

The tools that rearrange pages or output them as images, i.e. merging, splitting, deleting pages, rotating and exporting as JPG or PNG, are stricter than the compressor. They also reject PDFs with interactive forms or XFA forms, as well as PDF portfolios, embedded file attachments and optional layers. These rules do not apply to images to PDF, because there you select JPG or PNG files.

The compressor can process ordinary, unsigned forms. It counts the form fields but does not check whether every field works exactly as before afterwards. So test filled-in or fillable fields in the saved file.

File size

Each PDF, JPG or PNG file may be at most 50 MiB, which is 52,428,800 bytes. For a DOCX file in Word to PDF, the limit is 20 MiB, i.e. 20,971,520 bytes. For tools with multiple inputs, the limit for all files combined is 100 MiB, i.e. 104,857,600 bytes. You can find all exact values and the supported features collected under File and page limits.

Browser and requirements

To use the tools, you need a browser with JavaScript enabled. Compression additionally requires WebAssembly and Web Workers. For previews and image features, the browser must be able to render pages and decode the respective image formats.

If a tool doesn't start or stops

If a tool reports an error before the actual work begins, reload the page so that the browser fetches the program components again. If processing of a large file stops, close other tabs and programs and try again. If that doesn't help, a different browser with the features mentioned or a device with more memory may help.

Processing locally is not the same as working offline

The processing itself takes place in your browser. However, the website and its program components have to be loaded first, and the browser loads some of them only when you use a tool. Without an internet connection, the website therefore cannot be reliably opened or restarted. If you need to work entirely without a network, the guide Compress PDF without uploading compares the possible options. There are separate guides for programs installed on your computer: Reduce PDF size on Windows and Reduce PDF size on a Mac.