What is the PDF file format?
Files with a .pdf extension contain text or graphic documents stored in a format designed to provide portability between applications and platforms. It stores a document of fixed appearance - a precise digital printout that renders identically regardless of the device or software used to open it. The specification was originally developed in the 1990s by Adobe Systems, and it is based on the PostScript language. Since 2008 it has been an open format defined under ISO 32000 - previously it was a proprietary file format. The most recent revision, PDF 2.0, is specified in ISO 32000-2 (published 2017, revised 2020).
The current implementation allows placing within a .pdf document:
- Text and embedded fonts,
- Vector graphics and raster images,
- Annotations,
- Interactive forms,
- Layers,
- 3D objects,
- Multimedia content (including video)
A .pdf file can be encrypted and digitally signed, and it may include metadata and file attachments. It is a synthesis of the PostScript page-description language, a font embedding system that bundles typefaces so the document renders consistently on any machine, and a structured object storage system that merges all elements into a single binary file with optional data compression.
The file always begins with the header %PDF- followed by a version string (e.g., %PDF-1.7 or %PDF-2.0) at byte offset 0, and closes with the %%EOF marker - making .pdf files reliably identifiable by their magic bytes. Form data exported from PDF documents can be saved separately as .fdf or .xfdf files. Adobe Illustrator .ai files, which share PostScript heritage with PDF, are closely related and can often be opened by PDF-capable viewers.
Security & safety
RISK: MEDIUMPDF is not inert: the format can embed JavaScript, launch actions, embedded files and links, and over the years attackers have used malicious PDFs to exploit reader vulnerabilities or to phish (a fake invoice with a link to a credential-stealing site). Keep your PDF reader updated, disable JavaScript in the reader if you do not need it, and be wary of unexpected PDFs prompting you to click links or enable content. Encrypted PDFs and password prompts are normal for sensitive documents. Most everyday PDFs are safe, but treat attachments from unknown senders with caution.
Format details
in a nutshellPrograms that open PDF files
Technical details
deep spec| MIME type | application/pdf |
| Magic bytes | %PDF- (hex: 25 50 44 46 2D) at byte offset 0; file ends with %%EOF |
| File structure | Object-based binary format; cross-reference table (xref) or cross-reference stream maps object numbers to byte offsets within the file |
| Compression filters | FlateDecode (zlib/deflate), DCTDecode (JPEG), JPXDecode (JPEG 2000), JBIG2Decode, LZWDecode, and additional stream filters |
| Encryption | AES-128 and AES-256 (PDF 1.6+/1.7+); legacy RC4 40-bit and 128-bit also defined in the specification |
| Digital signatures | PKCS#7/CMS and PAdES signatures embedded in the document; supports both approval and certification signature types |
| Color spaces | DeviceRGB, DeviceCMYK, DeviceGray, ICC-based (embedded ICC profile), CIE Lab, and spot/separation colors |
| Font embedding | Type 1, TrueType, OpenType (CFF/glyph outlines), and CIDFont types can be fully or partially (subsetting) embedded |
| Linearization | Optional Fast Web View structure enabling page-at-a-time streaming download without loading the entire file first |
| Conformance subsets | PDF/A (archival, ISO 19005), PDF/X (prepress, ISO 15930), PDF/E (engineering), PDF/UA (accessibility, ISO 14289), PDF/VT (variable/transactional data) |
| Version signaling | Declared in the file header (e.g., %PDF-1.7 or %PDF-2.0); the document catalog may specify an equal or higher version |
| Page coordinate system | Default user space: 1/72 inch per unit, origin at bottom-left; maximum page size 14,400 x 14,400 inches (PDF 1.6+) |
| Tagged PDF (accessibility) | Optional structure tree with logical reading order, alt-text, and role mapping for screen readers and PDF/UA (ISO 14289) compliance |
| Transparency model | Alpha-compositing model introduced in PDF 1.4: supports blend modes, alpha channels, and soft masks for layered objects |
| JavaScript support | Acrobat JavaScript (ECMAScript subset) can be embedded for interactive forms, field validation, and document-level automation |
| Metadata | Document-level XMP metadata stream plus legacy Info dictionary; supports Dublin Core, PDF, and custom extension schemas |
| Released | 1993 (Adobe); ISO 32000-1 in 2008, ISO 32000-2 (PDF 2.0) in 2017/2020 |
| Open standard | Yes · royalty-free |
| Specification | www.iso.org |
PDF conversions
Community Q&A
asked by usersNo questions yet - be the first to ask about PDF files.