kreuzberg-dev/kreuzberg v4.2.13

on GitHub

What's Changed

Added

• WASM Office Format Support: DOCX, PPTX, RTF, reStructuredText, Org-mode, FictionBook, Typst, BibTeX, and Markdown extraction now available in the browser/WASM build
• Citation extraction for RIS, PubMed/MEDLINE, and EndNote XML formats
• JPEG 2000 and JBIG2 image decoding for OCR (pure Rust)
• Gzip archive extraction with bomb protection
• WASM integration tests and e2e fixtures for all new office formats
• Security limits for archive extraction pipeline

Fixed

• WASM Build: Fixed zstd-sys and tokio/mio compilation failures for wasm32-unknown-unknown when office feature enabled
• MIME type detection: case-insensitive validation, missing types synced with extractor registry
• YAML, Typst, Djot file format recognition
• Benchmark harness fixes for Go, Java, PHP, Elixir, WASM wrappers

Full Changelog: https://github.com/kreuzberg-dev/kreuzberg/blob/main/CHANGELOG.md#4213---2026-02-07