- Add official GNU Affero General Public License v3.0 to uvr_cli/LICENSE
- Add SPDX-License-Identifier: AGPL-3.0-or-later and Copyright (C) 2026 RevLayDust headers to all uvr_cli modules
- Document multi-license architecture in root README.md and uvr_cli/README.md
- Add automated regression tests in tests/test_uvr_cli.py verifying license file and module SPDX headers
- Implement modular 5-phase architecture (Detect -> Plan -> Verify -> Install -> Validate + Record) compatible with PowerShell 5.1
- Consolidate all dependency pins, thresholds, and URLs in Config.ps1 as single source of truth
- Pure, deterministic Plan generator supporting non-destructive DryRun
- Hardware and invariant verification gate (NVIDIA CC >= 8.0 for TRT-RTX, min driver 570, disk space, process locks)
- Safe mutation engine with pip freeze snapshot rollback and atomic install_record.json persistence
- Add Pester test suite covering 15 scenarios across planning, verification, and recording
- Convert install_packages.bat into thin CRLF launcher delegating to Install-UVR.ps1
- Fix CMD block parsing syntax error where literal parentheses inside if blocks closed the block prematurely
- Verified clean execution without syntax errors across both CUDA and CPU modes
- Add inspect-stack action in install/install_record.py to detect active Torch/ORT builds
- Detect stack transitions (CPU to GPU, GPU to CPU, Option A to B) when reusing venv
- Cleanly replace PyTorch packages with --reinstall-package to prevent uv skipping cross-build wheels
- Seamlessly swap ONNX Runtime packages (1.22.0 vs 1.24.4) and manage EP plugins
- Add unit tests in tests/test_tensorrt_rtx.py and preserve CRLF in install_packages.bat
- Implement TensorRT-RTX backend via onnxruntime-ep-nv-tensorrt-rtx plugin in engine/tensorrt_rtx/
- Support in-memory FP16 downcast for MDX-Net ONNX models with zero disk writes
- Implement automatic fallback chain: TensorRT-RTX -> CUDA -> CPU
- Add GUI execution provider selection with dynamic hardware gating
- Add CLI provider flags and model resolver support for trt_rtx
- Add install/requirements_trt_rtx.txt and stack profile tracking
- Add test suite in tests/test_tensorrt_rtx.py and test_uvr_cli.py
- Update run_uvr.bat to execute UVR.py via native CMD
- install/install_packages.bat: fix Visual C++ 2015-2022 x64 registry detection to avoid redundant downloads
- README.md: streamline TensorRT-RTX and Windows ML execution provider roadmap items
- __version__.py: bump BUILD_ID to RLUV_20261006_1949
Add :ensure_vcredist subroutine to install_packages.bat that:
- Detects VC++ 2015-2022 x64 via HKLM registry key (VisualStudio\14.0\VC\Runtimes\X64, version >= 14.40)
- Falls back to checking Programs uninstall registry key for older installs
- Downloads vc_redist.x64.exe from official Microsoft aka.ms permalink if missing
- Installs silently with /install /quiet /norestart, handles exit code 3010 gracefully
- Non-fatal: warns and continues if download or install fails
- Runs after Python dependencies, before FFmpeg/Rubberband
- Bundle pre-compiled diffq-0.2.4-cp311-cp311-win_amd64.whl in install/wheels/
- Pre-install diffq local binary wheel and pass --find-links in install_packages.bat
- Remove diffq from requirements.txt and requirements_cpu.txt to prevent PyPI sdist compilation failures on clean Windows environments lacking Visual Studio
- Ensure onnxruntime CPU and GPU packages do not collide by cleanly uninstalling conflicting variant before installation
- Update APP_NAME branding to 'RLD UVR-Opt' and clarify OS compatibility warnings in install_packages.bat
- Remove obsolete root-level install_packages.sh script and ignore install/install_packages.sh in .gitignore
- Add placeholder text file to preserve output/ directory tracking in git
- Restructure README around Time to Hello World and One-Click Installation via install_packages.bat
- Modernize manual installation instructions using Astral uv (uv venv, uv pip)
- Emphasize Windows 10/11 64-bit and ONNX MDX-Net priority
- Update roadmap with Windows ML and TensorRT Execution Providers
- Compress demo_uvr-opt.webp asset and clean up unused demo.svg
- Detect FFmpeg and rubberband-cli in system PATH or root application directory
- Automatically download and extract portable binaries to application root if missing
- Preserve all existing installer routines and maintain strict CRLF line endings
- Add RevLayDust avatar image and repository link alongside core developers in Help Guide credits
- Synchronize LICENSE_TEXT in constants.py with root LICENSE copyright
- Register responsive rld_credits_img in ImagePath
- Update application title and console banner to UVR-Opt / RLD UVR-Opt
- Set application version to v0.1.0-beta and patch identifier to RLUV_20260915_2010
- Temporarily disable Adaptive Chunk Scheduler in settings to prevent crashes
- Suppress official update notification prompt by syncing latest version to current patch
- Add RLD credits branding asset in gui_data/img/ and remove compiled UVR.pyc artifact
- Implement PEP 562 lazy loading in uvr_cli/__init__.py to reduce import time from 7.85s to 0.02s
- Defer heavy inference dependencies (torch, separate, onnxruntime, numpy) into CLI command handlers
- Reduce uvr_cli --help startup latency from ~8.98s to ~0.20s (~45x faster)
- Introduce InferenceResult tuple subclass in runner.py to isolate pure model execution timing
- Accurately measure and report pure inference elapsed time in uvr_cli run and benchmark engine
- Add unit tests for CLI startup backend isolation, PEP 562 exports, and InferenceResult structure
- Synchronize user-specified overlap, segment_size, and batch_size attributes onto HeadlessModelData
- Prioritize active overlap_mdx value when writing benchmark CSV rows and JSON summaries
- Support shorthand precision strings (fp32, fp16, bf16) in build_headless_model_data
- Add unit test coverage verifying custom bench parameters are captured in model data and CLI parsing