VoxSafe-Offline Transcription

VoxSafe is a privacy-first voice recorder with on-device transcription and automatic speaker identification. Everything runs locally — your recordings, transcripts, and speaker data never leave your device. KEY FEATURES • Automatic Speaker Diarization VoxSafe identifies and labels different speakers in a conversation automatically, using on-device pitch analysis. No cloud processing, no privacy compromise. Manual speaker switching is also supported when you want full control. • On-Device Transcription Choose between Apple's built-in Speech framework (fast, online) or Whisper (fully offline, higher accuracy). Both engines run entirely on your device. Switch engines anytime, with automatic fallback if your preferred engine is unavailable. • Lock-Screen Playback Controls Pause and resume recordings directly from your lock screen or Control Center. Now Playing info keeps you informed of recording status even when the app is in the background. • Background Transcription Transcription continues even if you switch away from the app. A background task keeps the process alive, and the status is tracked so you can see if a transcription was interrupted or is still pending. • Recording Recovery If the app is force-quit or the system kills it during a recording, VoxSafe automatically recovers the incomplete recording on next launch and fixes the audio header so the file remains playable. • Markers & Notes Drop timestamped markers during recording to highlight key moments. Add a quick note to each marker for later reference. • Flexible Export Export transcripts as SRT subtitles or CSV for use in other tools. Share full transcript text or individual segments. Share the original audio file anytime. • Import Audio Import existing audio files from your Files app or other sources. VoxSafe transcribes them just like a live recording. • Search History Full-text search across all transcripts. Find the exact conversation in seconds. • Editable Titles Rename each recording with a custom title at save time. Edit titles later from the detail screen. • Security Optional Face ID / Touch ID lock with PIN fallback keeps your recordings private. • Multi-Language Transcription supports Chinese, English, Japanese, and Korean. Interface available in Chinese and English. PRIVACY VoxSafe is designed for sensitive scenarios — legal consultations, medical notes, confidential interviews, and private meetings. All processing happens on-device. No accounts, no cloud uploads, no tracking. ENGINES • Apple Speech: Uses the system's built-in speech recognition. Fast and accurate for online use. • Whisper: OpenAI's open-source model running locally on your device. Fully offline, no network required. Download the model once (small, base, or medium) and transcribe anywhere.