Xberg (Kreuzberg) plugin
Document extraction engine for PDFs, Office files and scans, shipped as an agent plugin, MCP server and CLI
- 9,458 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Runs a native binary and reads project files
- Makes network requests when URL ingestion is enabled
Install
- Works in
Manual install
/plugin marketplace add xberg-io/xberg
/plugin install xberg@xbergAccess and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Runs a native binary and reads project files
- Makes network requests when URL ingestion is enabled
This is third-party code. Review the repository files before installing.
What it does
Xberg parses more than a hundred document formats: PDF, Word, Excel, slides, email, e-books, archives and images. It returns clean text or Markdown, tables, metadata and images, with OCR for scans via Tesseract, PaddleOCR or vision models. The plugin gives the agent skills for batch processing, OCR, table and keyword extraction, chunking and picking an output format. The same engine runs as an MCP server with xberg mcp, so the agent can call extraction directly instead of shelling out.
Who it is for. For RAG developers, analysts and anyone who needs an agent to read documents locally without cloud services.
Good fit when
- You need text and tables from a batch of PDFs or scans
- You are building RAG and need document parsing with chunking
- The agent writes Python, Node.js or Rust code that calls Xberg
Not a fit when
- You only need to read one plain text file
- You cannot install native binaries or Docker on the machine
Example request
Extract all tables from these PDF reports as Markdown and use OCR for the scanned onesLimitations
The MCP server needs the xberg CLI installed (brew, Scoop or Docker). Some features such as transcription and URL ingestion are behind build flags. LLM-based structured extraction needs a local model or a provider key. The project was recently renamed from Kreuzberg to Xberg, so older guides may use previous package names.
How to disable. Uninstall the xberg plugin via /plugin in Claude Code or the equivalent command in another agent, and remove the xberg block from your MCP config.
FAQ
How is Xberg related to Kreuzberg?
It is the next version of the same engine under a new name; the old line lives in a separate kreuzberg-v4-lts repository.
Does it need a GPU?
No, it runs on CPU by default.
Discussion
No questions or reviews yet.
Related
PluginsMCP PDF for large documents: the agent searches by meaning and keyword across one file or a folder and reads only the pages, tables and scans it needs
An MCP server built on Docling: converts PDFs into structured documents, builds new ones and feeds them to RAG
An MCP server that turns PDF, DOCX, XLSX, PPTX, images, audio, web pages and YouTube transcripts into Markdown
PageIndex MCP: long PDFs for an agent
PageIndex MCP
An MCP server that lets an agent read long PDFs through a document tree: retrieval by reasoning instead of a vector database
A PDF-reading MCP server: structured text, tables, OCR, visual crops and page-level citations
Microsoft's utility and MCP server that turn PDF, Word, Excel, PowerPoint, HTML and other files into Markdown for models
README in short
The README covers a Rust core with bindings for about fifteen languages and several ways to run it: library, CLI, REST API, MCP server, Docker and Helm. It lists supported formats, OCR backends, layout and table models, embeddings and text enrichment. A separate section explains installing the plugin in Claude Code, Codex, Cursor, Gemini CLI, Copilot CLI, Factory Droid and opencode. MIT licensed.