FlashVoice
Version History

Changelog

Stay up to date with the latest features, improvements, and fixes in FlashVoice.

v1.4.5

Latest
8/11/2026
New
  • FlashVoice is now open source. The source code is available on GitHub for review, auditing, and community contributions.
Improvements
  • Optimized core processing to improve startup, voice input, and file transcription performance.
  • Reduced resource usage during long-running sessions for better stability.
  • Improved UI responsiveness and overall interaction smoothness.
Fixes
  • Fixed multiple issues affecting voice input, file transcription, and keyboard shortcuts.
  • Corrected inconsistent states and UI display issues in several scenarios.
  • Resolved several stability issues that could appear during extended use.

v1.4.4

12/10/2025
Improvements
  • Faster startup and reduced latency in Push-to-Talk and Toggle modes.
  • Improved multilingual recognition accuracy (Chinese, English, Japanese, Korean).
  • Enhanced AI Polish performance for long documents.
  • Refined UI responsiveness for multi-monitor setups.
Fixes
  • Fixed occasional auto-insert failures when switching between apps.
  • Corrected progress display for large video transcription tasks.
  • Fixed hotkey inconsistencies on non-US keyboard layouts.

v1.4.3

11/22/2025
Improvements
  • More reliable Hands-free mode activation.
  • Smoother real-time transcription animation.
  • Better system audio capture performance under high CPU load.
Fixes
  • Resolved rare device switching crashes.
  • Fixed incorrect timestamp alignment in long sessions.

v1.4.2

11/5/2025
Improvements
  • Optimized language auto-detection for offline transcription.
  • Faster batch processing for large file queues.
  • Improved punctuation and phrasing rules in AI Polish.
Fixes
  • Fixed UI freezing in the file import panel.
  • Resolved incorrect SRT export formatting in mixed-language audio.

v1.4.1

10/18/2025
Improvements
  • Reduced RAM usage during local model loading.
  • Faster transitions between Push-to-Talk and Toggle mode.
  • Improved segmentation for overlapping speakers.
Fixes
  • Corrected cursor placement when inserting text into web editors.
  • Fixed issue where file transcription would occasionally pause indefinitely.

v1.4.0

9/30/2025
New
  • Hands-free continuous transcription mode.
  • Auto-detect language for imported audio/video files.
  • Support for more formats: FLAC, AAC, M4A, MP4.
  • New Quick Export panel (TXT / Markdown / SRT).
Improvements
  • More stable offline STT pipeline.
  • Better timing accuracy for SRT exports.
  • Enhanced UI contrast for dark mode.

v1.3.3

9/15/2025
Improvements
  • 25% improvement in real-time transcription latency.
  • Higher accuracy for rapid speech and technical vocabulary.
  • Polished Smart Polish outputs for formal writing.
Fixes
  • Fixed audio desync when switching from mic to system audio mid-session.
  • Resolved occasional stuck indicator in Toggle mode.

v1.3.2

8/28/2025
Improvements
  • Faster loading of local LLMs (Ollama, Qwen, DeepSeek, GLM, etc.).
  • Improved prompt preset system for writing style customization.
  • More stable performance when working with large documents.
Fixes
  • Fixed rare duplicate insertion during rapid key presses.
  • Corrected Korean spacing issues in long text outputs.

v1.3.1

8/12/2025
Improvements
  • Improved UI for hotkey customization.
  • Better separation of sentences in English and Chinese.
  • Reduced CPU load during long continuous sessions.
Fixes
  • Fixed intermittent failures in file drag-and-drop.
  • Resolved timestamp drift in multi-hour audio files.

v1.3.0

7/30/2025
New
  • Local LLM proofreading support (fully offline).
  • Writing-style presets for AI Polish.
  • Customizable hotkeys for input mode switching.
  • Expanded model provider support: OpenAI, Anthropic, Google, DeepSeek, Qwen, Kimi, etc.
Improvements
  • Smarter punctuation rules.
  • Improved offline language modeling for mixed-language speech.

v1.2.2

7/18/2025
Improvements
  • Faster queue processing for batch transcription.
  • Enhanced drag-and-drop experience and file preview.
  • Improved error recovery for corrupted audio files.
Fixes
  • Fixed stalled transcription for certain MP4 files.
  • Corrected Markdown export formatting.

v1.2.1

6/28/2025
Improvements
  • Better handling of long pauses and filler words.
  • Higher accuracy for Japanese and Korean offline STT.
  • Faster waveform rendering for imported recordings.
Fixes
  • Fixed inconsistent progress bar updates.
  • Resolved rare crash when canceling mid-transcription.

v1.2.0

6/12/2025
New
  • Offline audio/video transcription (MP3, WAV, MP4, etc.).
  • Batch transcription with SRT, TXT, and Markdown export.
  • Language auto-detection for imported media.
  • Korean UI and transcription support.
Improvements
  • More stable privacy mode (no network dependencies).
  • Lower latency for model loading and initialization.

v1.1.1

6/3/2025
Improvements
  • Faster Push-to-Talk activation.
  • Improved Smart Polish grammar correction.
  • Reduced CPU usage during long dictation sessions.
Fixes
  • Fixed insertion bugs in several macOS native fields.
  • Corrected cleanup of temporary audio buffers.

v1.1.0

5/20/2025
New
  • Smart Polish engine: Auto-punctuation, Grammar correction, Clarity improvements, Multiple AI provider support.
  • Local LLM proofreading for full offline use.
  • Pre-insertion preview for polished text.
Improvements
  • Improved background noise handling.
  • More responsive live transcription indicator.

v1.0.1

5/15/2025
Improvements
  • Better transcription accuracy in noisy environments.
  • Faster hotkey detection.
  • Improved system audio capture stability.
Fixes
  • Fixed timestamp drift for long recordings.

v1.0.0

5/10/2025
New
  • Push-to-Talk voice input anywhere you type.
  • Real-time offline transcription with instant insertion.
  • System audio + microphone capture.
  • Local processing only — no uploads, no cloud usage.
  • Supports Chinese, English, Japanese, and Korean.
  • Up to 5× faster than typing.