Skip to content

2026

  • 1 min read

1.11.1

This release makes voice-driven workflows faster, more reliable, and easier to set up.

Highlights

  • New Whisper Large v3 Turbo model: Download and use a faster local transcription model that delivers nearly the accuracy of Whisper Large while using about 1.5 GB of memory.
  • Continuous listening timeout: Set how long monitor-based voice commands should remain active without a matched command, from 1 to 60 minutes. Recognized commands automatically restart the timer.

Improvements

  • More resilient voice monitoring: A temporary speech-recognition error no longer ends continuous listening. Dopdock keeps monitoring for the next command and safely recovers if the audio stream stops unexpectedly.
  • Better long-audio handling: Overly long ambient-audio chunks are skipped so they do not delay recognition of short voice commands.
  • Smoother accessibility setup: The onboarding screen now detects permission changes automatically after you grant access in System Settings.

Fixes

  • Window focus: Switching Dopdock between background and foreground window modes no longer unexpectedly steals focus from the app you are using.
  • 1 min read

1.11.0

This release makes voice-driven workflows faster, more reliable, and easier to set up.

Highlights

  • New Whisper Large v3 Turbo model: Download and use a faster local transcription model that delivers nearly the accuracy of Whisper Large while using about 1.5 GB of memory.
  • Continuous listening timeout: Set how long monitor-based voice commands should remain active without a matched command, from 1 to 60 minutes. Recognized commands automatically restart the timer.

Improvements

  • More resilient voice monitoring: A temporary speech-recognition error no longer ends continuous listening. Dopdock keeps monitoring for the next command and safely recovers if the audio stream stops unexpectedly.
  • Better long-audio handling: Overly long ambient-audio chunks are skipped so they do not delay recognition of short voice commands.
  • Smoother accessibility setup: The onboarding screen now detects permission changes automatically after you grant access in System Settings.

Fixes

  • Window focus: Switching Dopdock between background and foreground window modes no longer unexpectedly steals focus from the app you are using.
  • 1 min read

1.10.2

  • Start a new dictation while an earlier one is still being processed, so your ideas keep flowing without waiting.
  • Keep dictation results in the order you recorded them, even when some take longer to process.
  • Get clearer status-bar controls while dictations are queued.
  • Improved compatibility and reliability with the Speechall transcription service.
  • 1 min read

1.10.1

  • Updated the Speechall Swift SDK to the latest pinned revision.
  • Improved compatibility with Speechall’s updated model-listing API and transcription schemas.
  • Speechall model IDs are now handled dynamically, supporting newly available models without requiring an app update.
  • 1 min read

1.9.5

  • Added experimental Speechall streaming transcription for faster dictation results while recording.
  • Kept Speechall streaming off by default and added an Experimental Features setting to enable it.
  • Improved recovery when a streaming transcription request fails so dictation can be started again without restarting Dopdock.
  • Improved local purchased-mode development launch logging.
  • 1 min read

1.8.2

  • Added Markdown mode for Paste Rich Text actions, including a plain-text Markdown editor and backward-compatible rich text actions
  • Improved rich Markdown paste behavior for Slack and other editors by adding HTML, RTF, and plain-text pasteboard formats
  • Improved Markdown code rendering for inline code and fenced code blocks
  • Added a debug Markdown paste action to make rich paste behavior easier to test
  • Added a local debug purchased launch script for testing purchased app behavior
  • 1 min read

1.8.1

  • Improved Markdown rendering for AI and action output, including better support for headings, lists, links, code blocks, quotes, bold, italic, and strikethrough text
  • Simplified the custom LLM model picker by removing the duplicate provider chip while keeping provider names in custom model titles
  • 1 min read

1.8.0

  • Added per-context dictation insertion overrides
  • Fixed GPT post-process toggle reset behavior
  • Improved dock icon handling for auxiliary windows
  • Show provider names for custom LLM models