NEW IN VERSION 1.0.23 ALIBABA CLOUD BAILIAN ASR SUPPORT • New STT provider: Alibaba Cloud Bailian (通义千问 ASR) • Powered by Qwen3-ASR-Flash model for superior accuracy • OpenAI-compatible API format for seamless integration • Ideal for users in China and Asia-Pacific regions • Enhanced multilingual recognition for Chinese, English, and more PROVIDER OPTIMIZATION • Removed Zhipu AI provider (30-second limitation) • Streamlined STT provider selection for better user experience • Smart audio format selection based on provider preferences • Improved compatibility and performance across all providers ENHANCED AUDIO TESTING • New multilingual sample audio files for testing • Smart format selection (M4A for OpenAI/Groq, WAV for Alibaba Cloud) • Better validation for your STT configuration COMPREHENSIVE STT PROVIDER OPTIONS Velox now supports 4 major STT providers to fit your needs: • OpenAI Whisper – Industry-leading accuracy and multilingual support • Groq – Lightning-fast transcription with generous free tier • Alibaba Cloud Bailian – Optimized for Chinese and APAC users • Custom Endpoints – Use your own Whisper-compatible service or local deployment RECENTLY ADDED FEATURES (v1.0.21-1.0.22) • Advanced Analytics Dashboard with interactive charts - Daily/Weekly/Monthly usage trends - Professional statistics cards (Total, Average, Peak) - Mini trend visualization on dashboard • SQLite Database Migration - Enhanced metadata support (app context, language detection) - Persistent statistics that survive history clearing - Improved performance and data integrity - Seamless auto-migration from UserDefaults DASHBOARD & STATISTICS (v1.0.19) • Characters Transcribed – See your total transcribed characters (displayed as 12.5K, 1.2M, etc.) • Days with Velox – Track how long you've been using Velox since first launch • Total Recordings and AI Refined counts at a glance 7-LANGUAGE SUPPORT • Full interface available in 7 languages: - English, 简体中文, 繁體中文, 日本語, 한국어, Deutsch, Español • Dual-track language system: UI language and AI output language can be set independently • Use English interface while getting Traditional Chinese output – perfect for multilingual users! APPEARANCE MODE • Independent appearance settings: - Light Mode – Always use light appearance - Follow System – Match your macOS setting (default) - Dark Mode – Always use dark appearance • Works with all theme colors (Flamingo, Ocean, Lavender, etc.) PRIVACY-FIRST DESIGN • No proprietary servers or hidden relays • API keys stored securely in macOS Keychain • Temporary audio files deleted after each run • Local history only, never uploaded • Full control over your data and API choices Questions or feedback? Email us at admin@wuhu-manxiu.com

NEW IN VERSION 1.0.21 DASHBOARD & STATISTICS • New Dashboard with usage statistics to track your voice input journey • Characters Transcribed – See your total transcribed characters (displayed as 12.5K, 1.2M, etc.) • Days with Velox – Track how long you've been using Velox since first launch • Total Recordings and AI Refined counts at a glance 7-LANGUAGE SUPPORT (Expanded!) • Full interface now available in 7 languages: - English, 简体中文, 繁體中文, 日本語, 한국어, Deutsch, Español • Dual-track language system: UI language and AI output language can be set independently • Use English interface while getting Traditional Chinese output – perfect for multilingual users. SIMPLIFIED VS TRADITIONAL CHINESE • AI output now supports separate Simplified and Traditional Chinese options • Solved the long-standing Whisper API limitation through advanced prompt engineering • Automatic detection: If your system is set to zh-TW/HK/MO, Traditional Chinese is used automatically APPEARANCE MODE • New independent appearance settings: - Light Mode – Always use light appearance - Follow System – Match your macOS setting (default) - Dark Mode – Always use dark appearance • Works with all theme colors (Flamingo, Ocean, Lavender, etc.) QUICK SETTINGS IMPROVEMENTS • Writing style selector now automatically disables when AI Refinement is off • Clear visual feedback with reduced opacity when disabled • Smarter UI interactions UI POLISH • Streamlined menu bar (removed redundant About item) • Fixed card width consistency in Settings • Improved StatCard layout WHAT'S NEXT We're constantly improving Velox based on your feedback. Thank you for your support! Questions or feedback? Email us at admin@wuhu-manxiu.com

What's New in Version 1.0.13: Major Improvements - Enhanced keyboard monitoring: ESC key cancellation now works reliably during transcription and refinement - Fixed paste functionality: text now pastes correctly after voice input completes - Improved hotkey detection: reduced false triggers when using Option+Backspace in terminal Technical Optimizations - Multi-layered keyboard listening strategy for better stability - Smart focus management for smoother user experience - On-demand health checks: only active when you're using the app - Graceful degradation: core features work even without accessibility permissions Other Improvements - User-friendly error messages with helpful guidance - Swift 6 concurrency improvements for better performance - Performance optimizations and bug fixes Thank you for using Velox! Your feedback helps us improve.

MAJOR UPDATE: Custom STT Provider Support NEW FEATURES: • Custom STT Provider Support – Now you can use ANY Whisper-compatible STT service! Configure your own Base URL, API Key, and Model name in Settings -> Speech to Text. This gives you complete flexibility and control over your voice input experience. • Local Whisper Deployment – Deploy Whisper locally for completely offline, privacy-first transcription. Perfect for users in China or those with strict privacy requirements. Your audio never leaves your device! • Flexible Service Options – Choose from: - Preset: OpenAI Whisper, Groq (super fast & free) - Local: faster-whisper-server, whisper.cpp, official Whisper - Self-hosted: Your own Whisper deployment - Any Whisper API-compatible service • Enhanced Provider Selection – Clear descriptions for each provider, including speed, accuracy, and cost information to help you make the best choice. IMPROVEMENTS: • Redesigned STT configuration UI with intuitive provider selection • Better error messages with actionable troubleshooting suggestions • Improved connection testing for custom services with detailed diagnostics • Enhanced model selection with smart defaults for each provider • Updated comprehensive documentation with step-by-step setup guides FOR CHINA USERS (中国用户专属): This update makes Velox AI Input fully usable in mainland China! You can now: - Use custom Whisper-compatible APIs (no international services required) - Configure proxy settings (Settings → Network) for international access - Deploy Whisper locally for 100% offline operation (完全离线) - No more dependency on blocked services – you're in full control! PERFORMANCE & STABILITY: • Optimized audio recording pipeline for better quality • Improved clipboard restoration reliability • Enhanced error handling for network timeouts and connection issues • Better memory management for long recording sessions • Faster app launch time DOCUMENTATION: • New comprehensive Custom STT Setup Guide • Local Whisper deployment tutorials • Troubleshooting guide for common issues • Best practices for different use cases BUG FIXES: • Fixed audio recording issues on some MacBook models • Resolved clipboard restoration failures in certain apps • Fixed proxy configuration not applying correctly • Improved error handling for invalid API responses • Fixed UI glitches in settings panel WHAT'S NEXT: We're constantly improving Velox AI Input based on your feedback. Coming soon: - More language support for UI - Advanced audio preprocessing options - Batch transcription support Thank you for using Velox AI Input! If you enjoy the app, please leave a review Questions or feedback? Email us at admin@wuhu-manxiu.com