Kreuzberg v4.0.0 introduces a complete rewrite in Rust, delivering enhanced performance and expanded language support for document intelligence workflows. The updated library now offers bindings for 10 languages including Python, TypeScript, Java, and Go, enabling consistent extraction across diverse tech stacks. Key improvements include removal of Pandoc dependencies, native Rust parsers, and a plugin system supporting custom OCR backends like Tesseract and EasyOCR.

New features such as ONNX embeddings, byte-accurate offsets, and streaming parsers make it ideal for LLM and RAG pipelines. Designed for production environments, Kreuzberg v4 includes REST API support, Docker images, and async-first architecture. Its MIT license ensures open access for integration into proprietary systems without legal constraints.

This release significantly benefits developers and enterprises relying on structured data extraction from 56+ document formats. By eliminating language barriers and boosting efficiency, Kreuzberg v4 strengthens its position as a foundational tool for AI-driven document processing.