TypeWhisper/typewhisper-mac v1.6.0-rc2

on GitHub

TypeWhisper 1.6.0-rc2 refreshes the 1.6 release-candidate line after rc1 with richer calendar-aware meeting recording, faster and more stable audio capture, expanded automation integrations, and a broad reliability pass across Recorder, local models, and cloud providers.

Highlights

• Improved recording startup performance in controlled hardware tests: median request-to-first-buffer time fell by 22.6% on the built-in MacBook microphone and 13.1% on a tested USB microphone, while opt-in AirPods recording readiness improved by about 45%.
• Added optional calendar-aware meeting automation with notifications, countdown controls, meeting activity checks, and automatic Recorder sessions.
• Calendar-started recordings now use sortable date-and-title filenames and retain the event title, time, location, organizer, and attendees in a versioned transcript sidecar, including when transcription is disabled or retried later.
• Improved Whisper Mode amplification with smoother gain changes, less amplification of low-level background noise, stable handling of brief pauses, and protection against clipping; standard recording mode remains unchanged.
• Added an MCP Client action add-on with Streamable HTTP support and refreshed tool discovery.
• Expanded the Obsidian integration with live sync and Markdown templates.
• Added bulk plugin updates with automatic restart after compatible updates are installed.
• Added Simplified Chinese localization across the app and bundled integrations; two AirPods keep-alive descriptions remain in English.

Dictation and Recorder Reliability

• Moved recording startup work off the main actor and prewarmed eligible microphone inputs to reduce the delay before audio capture begins.
• Added live transcript updates to active text fields while preserving the final insertion path.
• Fixed automatic meeting transcripts to use the finalized recording as their source of truth, with a captured-audio fallback when the saved file cannot be read.
• Added recovery-audio retention controls and improved Recorder retranscription when local models have unloaded.
• Fixed auto-submit in apps that do not expose standard Accessibility text roles.

Models and Integrations

• Expanded OpenAI-Compatible profiles with optional API versions, realtime transcription, and Responses API support.
• Fixed Qwen3 long-audio cutoff and loop-fallback behavior.
• Improved WhisperKit model restore after unloads and for workflow-specific model overrides.
• Improved Soniox cleanup and quota reporting, and fixed Speechmatics US batch routing and language-pack handling.
• Fixed ElevenLabs validation for restricted Speech-to-Text API keys and now surface specific permission and network errors.
• Tightened readiness handling for restorable transcription models.

UI and Platform Fixes

• Fixed calendar permission requests and the calendar entitlement used by distributed builds.
• Fixed the Settings sidebar scroll pocket on macOS 27 while preserving the existing layout on older macOS versions.
• Added regression coverage for TextEdit correction learning and the new meeting-recording final-transcription path.

Known Limitation

• Automatic private iCloud sync remains disabled while the production container and Developer ID provisioning profile are unavailable.
• Cloud Folder Sync through a user-selected iCloud Drive, Dropbox, OneDrive, Syncthing, or custom folder remains available.

Testing Focus

• Verify an upgrade from 1.6.0-rc1 and a clean installation, including permissions, first dictation, preserved user data, and the Release Candidate update channel.
• Exercise microphone startup and route recovery with built-in, AirPods/Bluetooth, and virtual inputs, including rapid start/stop and permission recovery; in Whisper Mode, cover quiet and loud speech, background audio, brief pauses, and clipping protection.
• Verify calendar meeting notifications, countdown actions, automatic recording, sortable filenames, finalized Recorder files, complete final transcripts, and .transcript.json event metadata with transcription enabled, disabled, failed, and retried.
• Smoke-test MCP Client actions over stdio and Streamable HTTP, Obsidian live sync and templates, bulk plugin updates, and automatic restart.
• Exercise OpenAI-Compatible realtime transcription and Responses API profiles, Qwen3 long audio, WhisperKit restore, Soniox quota errors, restricted ElevenLabs keys, and Speechmatics US batch transcription.
• Verify Simplified Chinese localization and Settings behavior on macOS 27 as well as the retained macOS 14-26 layout.
• Confirm v1.6.0-rc2 publishes only as a GitHub prerelease on the release-candidate Sparkle channel and does not update Homebrew or stable website messaging.