New Features - You can now manually send transcripts, summaries, and minutes to Webhook / Slack / Google Drive from the detail screen. - The send menu only shows services that are enabled. - Added recording_id to the Webhook payload. Improvements - Improved overall usability in the Settings screen. - Fixed an issue where changes were delayed on the detail screen.

New Features - Added "type" to the webhook payload - Added webhook payload preview Example of the JSON that will be sent: { "text": "(Body)", "type": "transcript", "recorded_at": "2026-10-09T10:00:00Z", "sent_at": "2026-10-09T10:05:00Z", "source": "whisperdirect", "version": "1.0" } Improvements - Webhook now sends transcriptions, summaries, and minutes at each completion stage - Localized some UI strings - Improved accuracy and responsiveness of live transcription (now uses the new Apple Speech engine on iOS 26+) - Fixed timestamp display in live transcription - Improved visibility of the Retranscribe button in the segment editor

New Features - Live Preview (real-time transcription) You can now see your speech transcribed in real time while recording. - Segment Re-transcription You can now re-transcribe individual segments in the transcription editor to fix misrecognized parts. (Uses OpenAI gpt-transcribe, billed per usage.) - Re-transcribe with Whisper API Even when using Apple Speech, you can re-transcribe specific items with the Whisper API. (Menu: "Re-transcribe with Whisper API") Design Updates - Display mode switching on the recording screen Choose from three modes via the dropdown at the top right of the header while recording. ・Live Preview: Shows real-time transcription ・Waveform (default): The conventional waveform display ・Simple: No waveform, just a blinking REC indicator Your selection is saved and retained for the next recording. - Header information consolidated: Recording status and elapsed time are now unified on the left side of the header, like "●REC 00:01" - Simplified bottom area: Only the waveform (Waveform mode only) and microphone name are displayed Improvements - Added a confirmation dialog when closing the transcription editor with unsaved changes - The edit button is now disabled while transcription or diarization is running Performance Improvements - Optimized waveform rendering: Reduced audio processing load and drawing overhead - Reduced CPU usage when idle: Waveform updates stop when not recording - Stable even for long recordings: Designed to keep memory usage constant

New Features - Added support for Gemini, now available for summary and minutes generation. - Added support for custom providers (OpenAI-compatible APIs). - Redesigned the API settings screen to manage OpenAI / Gemini / Custom in one place. Improvements - Optimized recording and playback to reduce CPU usage and battery consumption. - Improved diarization display so the first segment is correctly highlighted right after playback starts. - Reorganized the usage screen as "OpenAI API Usage." Design Adjustments - Made minor UI design adjustments. Other - Fixed minor bugs.

[NEW] ・Speaker diarization: NVIDIA Nemotron 3 runs on-device to automatically identify and color-code multiple speakers. ・Two transcription engines: switch between Apple Speech (on-device, free) and Whisper API (OpenAI, high-accuracy) in settings. ・Webhook integration: send results to Discord, Slack, n8n, and more by simply setting a URL. Header auth (API key) is also supported. ・Redo transcription: re-run transcription with the current settings. ・Change language and re-run: switch the language from the detail view to re-process automatically. [IMPROVED] ・Reduced battery consumption during recording (optimized waveform and timer rendering). ・Free trial changed from "5 runs" to "7-day trial". ・Added GPT-6 Luna to the LLM model options for summaries and minutes. ・Refreshed the onboarding screens. ・Cleaned up the settings screen UI. ・Refreshed the paywall screen. [OTHER] ・Added a new-version notification on launch. ・Show "What's New" after an update.

- Improved the appearance of the list screen

- Screen turns black while recording - Locked screen orientation to portrait on iPhone

- Redesigned the app icon. - Performance improvements and stability enhancements.

• Expanded Store Localization: Added localized App Store pages for German, French, Korean, and Chinese to reach more users worldwide. Note: The app interface is currently available in English and Japanese.

• Made internal improvements and minor adjustments.

• Bluetooth microphone support is now available. You can easily switch between microphones using the selector above the record button.

• Adjusted the community link behavior in the Settings screen — it now opens via an HTML redirect for improved compatibility.

• Added a link to the community thread on X in the Settings screen. You can now join the official community directly from the app.

• bugfix(mac): import screen not opening

• Added OCR transcription from images. You can now process multiple images at once, making it easy to convert documents and screenshots into text. All processing is done locally, so no API usage fees apply.

• Added iPad-style layout on Mac for improved usability on large screens. • Fixed an issue where files imported via Share could be saved twice. • Introduced automatic cleanup of temporary files at startup. • Other minor bug fixes.

• Fixed crashes when running on macOS.

• Fixed an issue where the internal FFmpeg update in the previous release caused import errors with certain formats.

The payment system has been updated, and trial usage is now available.

•Fixed a minor bug where the Slack test message was always sent in Japanese.

•Support for iOS 26 •Supports backup to Google Drive. •Added new LLM models: gpt-5-mini / gpt-5-nano •Adjusted default prompts (To apply the latest prompt, go to Settings > Summary Prompt / Minutes Prompt > Reset to Default)

Updated App Store screenshots.

•Summaries and meeting minutes now support the mini model for higher accuracy •Design improvements •Minor bug fixes and stability improvements

•Added Slack integration Now you can automatically post transcriptions, summaries, and meeting minutes to your selected Slack channel. •Minor bug fixes and usability improvements (Improved summary/meeting minutes button behavior, adjusted settings screen layout, etc.)

- Improved default prompts for minutes and summaries, and fixed an issue where generation could sometimes fail - Minutes and summaries can now be exported in Markdown (.md) format - Improved the generation button for minutes and summaries; re-generation is now possible