VoiceScribe: Voice to Text AI

Record a lecture, meeting, or voice memo — and get a full transcript plus a bullet-point summary, generated entirely on your phone. No cloud. No account. No minute limits. Works in airplane mode, straight from install — the AI model comes with the app, so there's nothing to download before your first recording.

UNLIMITED FREE TRANSCRIPTION
Other transcription apps cap you at a few hundred minutes a month because they pay for cloud servers. VoiceScribe runs its AI speech-to-text on your device, so transcription is unlimited and free — forever.

100% PRIVATE, 100% OFFLINE
Your audio never leaves your phone. There are no accounts, no analytics, and no uploads. Record confidential meetings, medical notes, or personal memos with confidence — everything stays in your pocket.

WHAT YOU GET
• Voice recorder tuned for speech — keeps recording with the screen off
• Automatic transcript with tap-to-play: tap any sentence to hear that moment
• Automatic bullet-point summary of every recording
• 99 languages with automatic language detection
• Import audio files (MP3, M4A, WAV, OGG, FLAC) — unlimited
• Recordings auto-title themselves from what was said
• Audio playback at 1x, 1.5x, or 2x speed
• Export transcripts as PDF or Word (.docx)
• Speaker labels for meetings and interviews (English)
• Share transcripts and summaries to WhatsApp, email, or any app
• Works the moment you install it — the AI model is included, no setup download

EVERYTHING IS FREE, AND OPEN SOURCE
No in-app purchases, no subscriptions, no ads, no locked features. Every feature listed above is free. The complete source code is public, so anyone can verify the privacy claims on this page: github.com/llassan/voice_scribe

PERFECT FOR
• Students recording lectures and turning them into study notes
• Meetings and interviews — get minutes without typing
• Journalists, researchers, doctors, and lawyers who need privacy
• Voice memos, ideas, and to-do lists captured hands-free
• Transcribing old audio files you already have

HOW IT WORKS
VoiceScribe uses Whisper-class AI models running directly on your phone's processor. The fast model is built in and ready the first time you open the app. If you record longer lectures or meetings, you can download a larger, more accurate model — entirely optional, and the built-in one keeps working either way.

Note: transcription speed depends on your phone. Longer recordings take longer to process on older devices — but they always finish, even with the app in the background.

Built on whisper.cpp and OpenAI Whisper (both MIT licensed).

Voice recorder, speech to text, audio to text, transcribe, meeting notes, lecture recorder — all in one private app.