Transcriptinator

Transcriptinator turns your voice into structured notes without sending audio anywhere. RECORD & TRANSCRIBE, FULLY OFFLINE Tap Record. Speak. Get a transcript. Everything happens on your iPhone using open-source Whisper models — audio never leaves your device, and the app works with no internet after initial setup. LEARNS WHO'S SPEAKING Enroll your voice once and Transcriptinator recognises you in every future recording. Add colleagues, family, or interview subjects with one guided sample each. Multi-speaker conversations get labelled automatically. IMPORT EXISTING AUDIO Bring in voice memos, meeting recordings, or any audio file from Files. Transcriptinator decodes WAV, M4A, MP3, AIFF, and more, then processes them the same way as live recordings. PICK YOUR QUALITY TIER - Fast (Whisper base, ~200 MB): quick turnaround for short notes - Balanced (Whisper small, ~600 MB): the everyday default - Accurate (Whisper medium, ~1.4 GB): best quality for interviews and meetings OPTIONAL: BRING YOUR OWN AI Connect your own API key from Anthropic, OpenAI, or Google to unlock: - Automatic document classification (meeting, todo list, journal, interview, etc.) - Todo extraction with assignees and due dates - Calendar event extraction with dates and locations - Question, decision, and note extraction - Custom extraction rules you write yourself Only text transcripts are sent to the provider you choose. Audio and speaker embeddings never leave your device. The app is fully useful without any AI configured. WHAT YOU'LL SEE - Home: your recordings, tap to play, swipe to transcribe or delete - Recording detail: full transcript with per-speaker chips and tap-to-seek - People: manage voice profiles, rename, merge, or delete - Extractions: todos, events, and notes with one-tap "Add to Calendar" or Share WHY OFFLINE MATTERS - No monthly subscription for transcription - Works on planes, in remote locations, at conferences with poor WiFi - Sensitive conversations (medical, legal, personal) stay on your phone - No account required — install and go Requires ~1 GB of storage for the recommended model bundle plus your recordings. Downloads happen once during setup.