OCRmyPDF Releases

This release focuses on improving debugging capabilities and network observability for the Uncloud CLI and daemon. It introduces new troubleshooting commands for streaming machine logs and measuring round-trip time (RTT), enhances log visibility during deployment failures, and adds a nightly release channel for testing updates.

Apr 24, 2026

This release introduces pre-deploy hooks for running setup tasks like database migrations and redesigns the deployment plan output for better readability. It also updates the default SSH connection method to use the system CLI and adds support for Compose file extensions like `x-context`, `pid`, and `shm_size`.

Apr 10, 2026

This release fixes text extraction issues for Right-to-Left languages and a bug where the working folder was not set in the Python API options.

Apr 6, 2026

This release adds a WebSocket generator for real-time LLM security testing and supports OpenAI Harmony format for audio inputs. It also introduces a probe tier for listing and improves reporting capabilities, while removing deprecated nemollm generators.

Apr 3, 2026

Apr 1, 2026

This release focuses on updates for Difftastic, a structural diff tool. It includes fixes for Git support to prevent crashes on unmerged paths, updates to parsers for multiple programming languages, and improvements to the build process and CLI error reporting.

Mar 16, 2026

This update refactors the Audio2Face-3D sample client configuration to version 2.0, incorporating new diffusion model settings and updated regression models. It also expands the blendshape library with 17 new tongue shapes.

Mar 11, 2026

Mar 8, 2026

This release addresses critical bugs in the Python API and improves PDF OCR handling. Key fixes include correcting the language parameter handling, resolving the tesseract_timeout issue that caused empty OCR output, and fixing text layer displacement on PDFs with non-zero MediaBox origins.

Feb 21, 2026

This release introduces a new LoRA-based abliteration engine with 4-bit quantization support and adds memory management features. It also enhances optimization workflows by allowing progress saving/resuming and broadens support for Vision-Language models.

Feb 14, 2026

Feb 11, 2026

Feb 9, 2026

Major refactoring of the report generator, support for JSON configs, and removal of deprecated features like `--generate_autodan`.

Feb 4, 2026

Jan 31, 2026

A minor maintenance release focusing on fixing LaTeX command escaping bugs and restoring the Apache2 license.

Jan 31, 2026

Jan 30, 2026

Jan 30, 2026

Major platform upgrades to Python 3.14 and PyTorch 2.9, introduction of new inference engines (Fireworks, OpenRouter), and rule-based evaluation judges.

Jan 29, 2026

Jan 22, 2026

Dec 24, 2025