v2.114.0

safing/portmasterv2.114.0Jul 20, 2026by docling-ops[bot]

AI Summary

Introduces support for legacy binary office formats and video processing pipelines, alongside new ASR (Automatic Speech Recognition) presets. Fixes include improvements to PPTX parsing safety, JATS list structure preservation, and VLM metadata handling.

Key Highlights

  • Added support for legacy binary office formats
  • Added VideoPipeline and VideoFormatOption for video processing
  • Added English-only and Distil-Whisper presets to native ASR backend
  • Fixed PPTX parsing using a safe XML parser

New Features

  • Support for legacy binary office formats
  • Parse native Word charts as classified pictures with data
  • Add VideoPipeline, VideoFormatOption, and InputFormat.VIDEO
  • Add English-only and Distil-Whisper presets to native backend

Full Release Notes

### Feature

* Add support for legacy binary office formats ([#3804](https://github.com/docling-project/docling/issues/3804)) ([`d0bca90`](https://github.com/docling-project/docling/commit/d0bca90b5a0ae8054186e54196ffb7032875927e))
* **msword:** Parse native Word charts as classified pictures with data ([#3809](https://github.com/docling-project/docling/issues/3809)) ([`26ff0b2`](https://github.com/docling-project/docling/commit/26ff0b25a17b632ddd3aa60312381dce93463c8d))
* **video:** Add VideoPipeline, VideoFormatOption, and InputFormat.VIDEO ([#3768](https://github.com/docling-project/docling/issues/3768)) ([`6f0df37`](https://github.com/docling-project/docling/commit/6f0df37cd087f5b170d2161106359fa367b3171f))
* **asr:** Add English-only and Distil-Whisper presets to native backend ([#3741](https://github.com/docling-project/docling/issues/3741)) ([`39b67b6`](https://github.com/docling-project/docling/commit/39b67b6f2d1132323d97381ec0289183c75b0e55))

### Fix

* Fix the profiling of NemotronOcrModel ([#3816](https://github.com/docling-project/docling/issues/3816)) ([`ccd09fb`](https://github.com/docling-project/docling/commit/ccd09fbfea9fba4aac57495e50afb9be6a889d1f))
* **pptx:** Use a safe XML parser ([#3831](https://github.com/docling-project/docling/issues/3831)) ([`a71c447`](https://github.com/docling-project/docling/commit/a71c44789198988e6a55717a73ad199400a7524f))
* **jats:** Preserve nested list structure in list items ([#3619](https://github.com/docling-project/docling/issues/3619)) ([`43f56c4`](https://github.com/docling-project/docling/commit/43f56c40191da8ddea14236b7c07e271cf359829))
* Skip empty display formulas in the JATS backend instead of truncating the document ([#3813](https://github.com/docling-project/docling/issues/3813)) ([`320c517`](https://github.com/docling-project/docling/commit/320c517b9c61cafa503ef7b1f1d808b01b3e88c9))
* Fall back to default image size for non-numeric AsciiDoc dimensions ([#3814](https://github.com/docling-project/docling/issues/3814)) ([`e398d5a`](https://github.com/docling-project/docling/commit/e398d5a2c826d4cf2a373c96706929fed517e58b))
* **vlm:** Use each element's own page for force_backend_text ([#3807](https://github.com/docling-project/docling/issues/3807)) ([`469e1ad`](https://github.com/docling-project/docling/commit/469e1ad87b8e5f4df5189cebc2747627a5b2af88))
* **visualization:** Scale top y-edge by scale_y in draw_clusters ([#3790](https://github.com/docling-project/docling/issues/3790)) ([`0844a97`](https://github.com/docling-project/docling/commit/0844a9709adefa4abfa21aff1fcb6609648e80fb))
* Use generation_config from options in PictureDescriptionVlmEngineModel ([#3778](https://github.com/docling-project/docling/issues/3778)) ([`e548307`](https://github.com/docling-project/docling/commit/e548307e8d32bb197c787b1f2998a87bafbc743b))
* **docx:** Clean up DrawingML temp dir on conversion error ([#3797](https://github.com/docling-project/docling/issues/3797)) ([`9b5aabc`](https://github.com/docling-project/docling/commit/9b5aabc4cfa48faab72c43b3b3581ed053c85897))
* **uspto:** Detect Grant Full Text Data/XML v2.5 patents case-insensitively ([#3801](https://github.com/docling-project/docling/issues/3801)) ([`6fb35f9`](https://github.com/docling-project/docling/commit/6fb35f94625d1d54bad678e52be6d5fe2263f483))
* **pptx:** Remove ineffective subtitle label expression ([#3785](https://github.com/docling-project/docling/issues/3785)) ([`d278116`](https://github.com/docling-project/docling/commit/d27811650d02847e250e92ee1b16cd7b415651cd))