v3.3.0 v0.1.8
VectifyAI/PageIndexSep 21, 2026 → Sep 21, 20260 days between releases
What's New
Newer release updates Docker to Python 3.13 and adds parsing features; older release added network model catalog refresh.
New Features
- Support for `MARKITDOWN_CU_ENDPOINT` and `MARKITDOWN_DOCINTEL_ENDPOINT` environment variables.
- Support for short YouTube URLs.
- Preservation of strikethrough semantics from HTML, CSS, and DOCX.
Improvements
- Refactored `markitdown-ocr` plugin for better maintainability.
- Updated Docker images to Debian trixie (Python 3.13-slim-trixie).
- Enhanced CSV parsing to handle UTF-8 BOM, blank rows, and line endings.
- Improved support for specific formats including YouTube short URLs and IPython notebooks.
Bug Fixes
- Fixed Unicode decoding errors for long files and malformed styles.
- Fixed CSV parsing issues regarding UTF-8 BOM and blank rows.
- Fixed math macro rendering errors.
- Fixed percent-encoded Windows drive paths in file URIs.
Upgrade Notes
Docker image updated to Debian trixie with Python 3.13. New environment variables `MARKITDOWN_CU_ENDPOINT` and `MARKITDOWN_DOCINTEL_ENDPOINT` are now supported. Expect largely unchanged behavior for typical inputs.
The Two Releases
v3.3.0
Sep 21, 2026· 1 month ago
A feature release designed to improve model availability and discovery by refreshing the model catalogue over the network at the beginning of a scan.
Full release notesVersion 0.1.8
Sep 21, 2026· 1 month ago
This release is a maintenance update containing numerous bug fixes and patches to improve stability and handling of edge cases across various file formats like CSV, DOCX, and EPUB. The `markitdown-ocr` plugin was also refactored for better maintainability.
Full release notes