Features Overview Leave feedback

GroupDocs.Metadata for Python via .NET reads, edits, removes, and exports metadata across documents, images, audio, video, and many other formats — Microsoft Office, PDF, OpenDocument, images (JPEG, PNG, PSD, TIFF), audio (MP3, WAV), video, fonts, e-books, archives, and more. It exposes the major metadata standards (XMP, EXIF, IPTC, Image Resource Blocks, ID3) and format-specific properties through one unified, predicate-driven API. It runs entirely on-premise, requires no Microsoft Office installation, and ships as a pre-built wheel on Windows, Linux, and macOS.

See the full list of supported formats or browse the Developer Guide for runnable examples of every API surface.

Read every property in a file, or search for specific properties using a plain Python predicate (lambda p: ...) — no specification objects required. You can match on a property’s name, value, type, or tags.

• Find metadata properties — locate properties that match a predicate, across any format.
• Extract metadata — pull properties from the native, document, and standards packages.
• Get known property descriptors — discover the strongly-typed properties a package exposes.
• Work with interpreted values — read human-friendly, interpreted forms of raw property values.

Add new properties or update existing ones in a unified way. Predefined tags let you set common properties (author, creation date, title) consistently regardless of the underlying format.

• Set metadata properties — add or update every property matching a predicate.
• Add metadata — add missing properties (for example, a last-printed date) when the format supports them.

Remove particular properties that match a predicate, or strip every detected property in a single call — ideal for privacy and compliance workflows.

• Remove metadata properties — delete only the properties that match a predicate.
• Clean metadata — sanitize() removes every detected property at once.
• Remove metadata (advanced) — remove whole metadata packages from a file.

Read and write the metadata standards used by images and many other formats:

• EXIF — read, update, and remove EXIF tags in JPEG, TIFF, PNG, WEBP, PSD, and more.
• IPTC IIM — read and edit IPTC datasets in JPEG, TIFF, and PSD.
• XMP — read, update, and add custom XMP packages across formats.

Image Resource Blocks (IRB) and format-specific native packages are exposed through the same root metadata tree.

Manage audio tags through the root package: ID3v1, ID3v2, Lyrics3, and APE, plus MPEG audio technical properties. Read or edit titles, artists, albums, and other tag fields, then export or strip them as needed.

Read basic facts about a file without walking its full metadata tree — format, MIME type, page count, size, and encryption state — and detect a file’s format by its internal structure rather than its extension.

• Get document info — format, MIME type, page count, size, and encryption flag.

Office documents also expose inspection data such as user comments, form fields, hidden pages, revisions, digital signatures, and common statistics (word count, character count).

The Metadata constructor accepts a file path, a binary file-like object, or a URI, so you can load files from local disk, in-memory buffers, or cloud storage.

• Load from a local disk
• Load from a stream — open("file.jpg", "rb"), io.BytesIO(data), or bytes fetched from S3, Azure Blob, or requests.
• Load a file of a specific format
• Load a password-protected document

After editing, save back to the original source or to a new destination — a path or a stream.

• Save to a specified location
• Save to a stream
• Save to the original source

Note

Saving a modified file requires a license. Without a license, the library runs in evaluation mode — it reads only the first few properties of each package and save() raises an “Evaluation only” exception. See Evaluation Limitations and Licensing.

Export the metadata tree — or a filtered subset of properties — to a spreadsheet or a structured data file for indexing, reporting, or downstream processing.

• Export metadata properties — write to Excel, CSV, JSON, or XML via ExportManager.

GroupDocs.Metadata is a useful building block for AI document pipelines: extract metadata as structured data to enrich search indexes and LLM context. The groupdocs-metadata-net pip package ships an AGENTS.md file inside the wheel so AI coding assistants can discover the API surface automatically, and GroupDocs runs a public MCP server for on-demand documentation lookups. See Agents and LLM Integration for the full story.

No cloud calls, no outbound network traffic, no third-party software dependencies beyond what the OS already provides. The wheel is self-contained on Windows and ships its own native runtime libraries on Linux and macOS. See System Requirements for the short list of optional native packages (libgdiplus, fontconfig).