XMP

XMP (Extensible Metadata Platform) is an ISO standard (ISO 16684-1) for embedding structured metadata inside files using RDF (Resource Description Framework) serialized as XML. It is widely supported across image, PDF, audio, and video formats. XMP metadata is human-readable, extensible (you can define custom namespaces), and standardized enough that tools from different vendors can read and write the same metadata.

What XMP is

XMP is a serialization of RDF/XML that fits inside the native container of a file format. A JPEG image, for example, stores XMP in an APP1 marker segment. A PNG image embeds it in an iTXt chunk. A PDF uses a Metadata stream. The specifics vary by format, but the XMP content itself — the RDF/XML — remains standardized.

How the packet is embedded

Each file format has its own XMP placement rules. JPEG places XMP in APP1 segments (the same segment that holds EXIF and thumbnail data). PNG uses iTXt (ancillary text chunks). PDF writes XMP to the document Metadata stream. MP3 uses ID3v2 frames. WAV and FLAC use their own metadata blocks. These placements are standardized so tools like ExifTool can read and write XMP consistently across formats.

Why LLMind builds on XMP

XMP is already the universal standard for portable embedded metadata. Rather than inventing a new container, LLMind builds atop XMP by defining a new namespace — https://llmind.org/ns/1.0/ — with AI-focused schema. The XMP packet provides the transport; the namespace provides the vocabulary. Files carrying LRFS metadata also carry standard XMP (EXIF, IPTC, Dublin Core), and all metadata coexists harmoniously.