📋 Changelog
Release Notes
Version history of Universal Translator
Latest
Version 2.6.2
Version 2.6.2
- 🛡️ Iron Command Architecture — A new, strict directive system. Guarantees 100% target language fidelity and completely eliminates context drops.
- 🎯 Dedicated Linguistic Cores — Universality replaced by strict specialization. Advanced marketing copywriting is now reserved exclusively for our most powerful units.
Version 2.6.0
Version 2.6.0
- System Smart Cache — Translate once. Reuse everywhere. Intelligent caching reduces redundant work and speeds up the entire process.
- Vision AI — Generate product descriptions from images in seconds. Perfect for e-commerce, marketing, and content creation.
- Smart CPU Thread Mapping — Optimized performance without overload. Balanced, efficient, and stable resource allocation.
- Redesigned Modern Interface — Faster, cleaner, more intuitive workflow that makes managing your translations a breeze.
Version 2.1.6
Version 2.1.6
- Implemented intelligent XML parsing
- System automatically skips tags containing images, links, product IDs, and long Base64 strings
- PDF → DOCX conversion: Converts PDFs to Word format before translation, improving preservation of the document layout
- Improved text chunking: Long texts are now split using native C++ tokenizers.
Version 2.1.2
Version 2.1.2
- Added full support for XML file format translation
- Added PDF format support with built-in OCR (Beta)
Version 2.0.9
Version 2.0.9
- Major architectural upgrade: Replaced NVIDIA CUDA backend with the universal Vulkan API
- Integrated automated Vulkan Runtime deployment into the setup installer for a seamless out-of-the-box user experience
Version 2.0.8
Version 2.0.8
- Introduced Marketing Style — translate content into a highly persuasive, SEO-optimized format with a single click
- Improved task scheduling efficiency for faster processing on multi-core systems
Version 2.0.7
Version 2.0.7
- Improved multithreading performance — now fully utilizes all CPU cores
- Added keyword generation feature
- Optimized memory management to reduce RAM usage during large workloads
- Implemented text chunking to prevent model overload when processing long texts
Version 2.0.4
Version 2.0.4
- Added built-in Windows binaries to prevent missing file errors
- Introduced optional GPU acceleration (500%–2000% performance improvement, depending on the GPU)
- Added CPU offloading for systems with limited GPU VRAM
- Implemented multithreading support
- Improved processed text structure
Version 2.0.1
Version 2.0.1
- Minor GUI improvements and refinements
Version 2.0
Version 2.0
- First stable release featuring a completely redesigned GUI