Voice-to-text dictation app with local (Nvidia Parakeet/Whisper) and cloud models (BYOK). Privacy-first and available cross-platform.
JavaScript aianthropiccross-platformgeminigroqlinuxmacosnvidiaopen-sourceopenaiparakeetspeech-to-texttranscribewhisperwindows
chore: bump to v1.5.2 with changelog and README updates
Gabriel Steinf217e2debc27
diff --git a/CHANGELOG.md b/CHANGELOG.mdindex caa43922b..427b229bf 100644--- a/CHANGELOG.md+++ b/CHANGELOG.md@@ -7,6 +7,33 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0## [Unreleased]+## [1.5.2] - 2026-02-24++### Fixed++- **Reasoning Output**: Resolved empty output for Qwen3/GPT-OSS models by raising local inference minimum tokens from 100 to 512; fixed custom endpoint models misrouting by checking `reasoningProvider` setting before name heuristics+- **Google OAuth**: Added `newUserCallbackURL` to desktop Google OAuth flow for proper new user registration+- **Linux KDE Taskbar**: Prevented dictation panel from appearing in KDE taskbar+- **Intel Mac CI Builds**: Fixed binary architecture mismatch by installing x64 ffmpeg-static binary and preventing prebuild hooks from deleting x64 binaries on arm64 CI runners (#196)++## [1.5.1] - 2026-02-23++### Added++- **GPU-Accelerated Local Inference**: Vulkan (Windows/Linux) and Metal (macOS) support for llama-server with automatic CPU fallback and GPU status badge in the reasoning model selector+- **CUDA GPU Acceleration for Whisper**: NVIDIA GPU acceleration for local Whisper transcription with automatic GPU detection, upgrade banner for existing users, and shared download progress UI+- **On-Demand Vulkan Download**: Vulkan llama-server binary downloads on-demand when the user opts in, saving 40-46MB from the app installer++### Changed++- **Vulkan Llama-Server Architecture**: Switched from bundling the Vulkan binary to on-demand download into userData, mirroring the Whisper CUDA download pattern++### Fixed++- **macOS Paste Failure**: Replaced osascript-based accessibility check with Electron's native `isTrustedAccessibilityClient()` and fixed focus transfer using hide()+showInactive() instead of blur() on NSPanel (#313)+- **Windows Sherpa-onnx Extraction**: Fixed tar extraction failing on Windows due to GNU tar interpreting drive letter colons as remote host separators — now uses relative paths (#284)+- **macOS Auto-Update Architecture**: Detect Rosetta translation via `sysctl.proc_translated` so Apple Silicon users stuck on an x64 build from older releases self-heal to the native arm64 build on next update+## [1.5.0] - 2026-02-23### Addeddiff --git a/README.md b/README.mdindex 90fd74d71..065bf816c 100644--- a/README.md+++ b/README.md@@ -730,7 +730,7 @@ A: OpenWhispr supports 58 languages including English, Spanish, French, German,## Project Status-OpenWhispr is actively maintained and ready for production use. Current version: 1.5.0+OpenWhispr is actively maintained and ready for production use. Current version: 1.5.2- ✅ Core functionality complete- ✅ Cross-platform support (macOS, Windows, Linux)diff --git a/package-lock.json b/package-lock.jsonindex 32f1cee98..8632690c6 100644--- a/package-lock.json+++ b/package-lock.json@@ -1,12 +1,12 @@{"name": "open-whispr",- "version": "1.5.0",+ "version": "1.5.2","lockfileVersion": 3,"requires": true,"packages": {"": {"name": "open-whispr",- "version": "1.5.0",+ "version": "1.5.2","hasInstallScript": true,"license": "MIT","dependencies": {diff --git a/package.json b/package.jsonindex 49ace3aba..9a62e0e07 100644--- a/package.json+++ b/package.json@@ -1,6 +1,6 @@{"name": "open-whispr",- "version": "1.5.0",+ "version": "1.5.2","description": "A desktop dictation application using whisper.cpp for speech-to-text transcription","main": "main.js","private": true,