Skip to content

Voice-to-text dictation app with local (Nvidia Parakeet/Whisper) and cloud models (BYOK). Privacy-first and available cross-platform.

docs: update README and CHANGELOG for v1.6.1

Gabriel Stein1373b31d5e4b
diff --git a/CHANGELOG.md b/CHANGELOG.md
index 14df676b0..125f36a39 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -7,6 +7,30 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
## [Unreleased]
+## [1.6.1] - 2026-03-08
+
+### Added
+
+- **WebSocket Streaming for BYOK Dictation**: OpenAI Realtime API streaming now works for standard dictation mode (not just meetings), enabling real-time transcription for Bring Your Own Key users
+- **Unified Streaming Path**: Extended OpenAI Realtime WebSocket streaming to normal dictation, sharing the same streaming infrastructure as meeting transcription
+
+### Fixed
+
+- **Transcript Loss on Disconnect**: Commit audio buffer before closing WebSocket and wait for final transcript before closing, preventing lost transcriptions during disconnects
+- **Dictation IPC Callbacks**: Send plain strings from streaming IPC callbacks instead of objects, fixing downstream consumers
+- **Accessibility Permission Detection (macOS)**: Fix onboarding flow not detecting macOS accessibility permission correctly (#394)
+- **Custom Cloud Provider Classification**: Treat Custom Cloud endpoints as self-hosted rather than third-party (#384)
+- **Blocking `execSync` in Meeting Detection**: Replaced synchronous process detection with async alternative to prevent UI freezes on Windows
+- **BYOK Onboarding Override**: Guard BYOK override for signed-in users and fix missing deps during onboarding (#397)
+- **Windows Media Pause Toggle**: Check audio state before sending media key on Windows (#402)
+- **Linux Wayland Portal Permissions**: Set desktop name on Linux for Wayland portal permissions (#389)
+- **Chrome Sandbox Permissions (Linux)**: Set SUID bit on chrome-sandbox during deb/rpm install
+
+### Changed
+
+- Eliminated duplication and fixed style inconsistencies in dictation streaming helpers
+- Cleaned up meeting detection code after the Windows input fix
+
## [1.6.0] - 2026-03-06
### Added
diff --git a/README.md b/README.md
index e2b3844a9..399cbbd5f 100644
--- a/README.md
+++ b/README.md
@@ -51,6 +51,7 @@ This project is licensed under the MIT License - see the [LICENSE](LICENSE) file
- πŸ“… **Google Calendar Integration**: Connect multiple Google accounts, view upcoming meetings in the sidebar, and receive auto-detection prompts
- πŸŽ™οΈ **Live Meeting Transcription**: Record and transcribe meetings in real-time via OpenAI Realtime API with automatic meeting detection (Zoom, Teams, FaceTime)
+- ⚑ **WebSocket Streaming for BYOK**: Real-time OpenAI Realtime API streaming for standard dictation mode, not just meetings β€” unified streaming path for all transcription
- πŸ” **Smart Detection**: Combines process monitoring, sustained audio detection, and calendar awareness to detect meetings automatically
### Notes
@@ -844,7 +845,7 @@ A: No. As of v1.6.0, OpenWhispr uses NSEvent monitors instead of CGEvent taps, e
## Project Status
-OpenWhispr is actively maintained and ready for production use. Current version: 1.6.0
+OpenWhispr is actively maintained and ready for production use. Current version: 1.6.1
- βœ… Core dictation with local and cloud processing
- βœ… Cross-platform support (macOS, Windows, Linux)