VoiceDraft: Offline AI Notes

Stop paying $99/year just to transcribe and clean up your own voice. VoiceDraft is the 100% on-device AI voice-to-draft studio built for thinkers, founders, journalists, and privacy-conscious professionals. Whether you are brain-dumping ideas on a morning walk, capturing a confidential client meeting, or drafting an email between calls, VoiceDraft turns messy spoken thoughts into structured, publication-ready Markdown—entirely on your hardware, with zero cloud uploads. WHY VOICEDRAFT? • 100% Offline & Airplane-Mode Ready No accounts. No sign-ups. No cloud servers. Your audio and transcripts never leave your phone's local sandbox. Turn on Airplane Mode and watch real-time transcription and local LLM rewriting run at full speed. • Pay Once ($19.99), Own Forever Every cloud AI note-taker locks your thoughts behind a $10–$20/month subscription wall. Because VoiceDraft runs AI natively on your phone's processor, we have zero server bills—and we pass that savings to you with a single lifetime unlock. • From Messy Ramble to Polished Draft (On-Device LLM) Don't settle for raw, unpunctuated transcripts. Powered by local GGUF language models (Qwen 2.5 & Llama 3.2 via Metal GPU acceleration), VoiceDraft transforms your transcript with 1-tap AI workflows: - Smart Polish: Strip out "ums," repetitions, and false starts into crisp, readable paragraphs. - Core Summary: Distill long voice memos into executive bullet points. - Meeting Minutes: Organize discussions into clear takeaways and decisions. - Email Draft: Turn a 60-second voice ramble into a well-structured business email. - Custom Prompts: Create and save your own reusable AI instructions for any workflow. - Non-Destructive Undo: Preview streaming AI output in real time, auto-generate concise titles, and revert with 1-tap Undo anytime. • Action Items Straight to Apple Reminders Let local AI scan your note to extract actionable tasks, manage your checklist right inside the note, and export unfinished tasks directly into iOS Apple Reminders with a single tap. • Merge Multiple Voice Fragments into One Long Draft Captured three separate voice ideas throughout the day? Select multiple notes from your workspace, drag to reorder them chronologically, and merge them into a single cohesive document—while keeping your original notes 100% intact. • Dual-Engine Offline ASR & Model Hub - Instant Out-of-the-Box Transcription: Ships with SenseVoice Small (INT8) + Silero VAD for ultra-fast, real-time streaming transcription in English, Chinese, Japanese, Korean, and Cantonese. - Downloadable Models with Smart RAM Guard: Download Whisper Base or switch between Qwen 2.5 (0.5B / 1.5B / 3B) and Llama 3.2 1B. VoiceDraft automatically detects your device's physical RAM to recommend the optimal model and prevent crashes. • Built for Deep Work & Obsidian Workflows - Pure Local Markdown Sync: Every note automatically syncs as a clean, WYSIWYG .md file in your device's "VoiceDraft/Notes" folder—ready for Obsidian or Logseq without messy technical tags. - Variable-Speed Audio & Verbatim View: Scrub through original recordings at 0.8x to 2.0x speed, or open the Verbatim Transcript sheet to view and copy the untouched raw speech. - Built-in Organization: Filter by Draft, Meeting, Interview, or Memo; swipe right to Pin; swipe left to Delete; or archive completed projects. FREE TIER VS. LIFETIME PRO • Free Forever Tier: Create up to 10 active notes (up to 3 minutes per recording) with full access to offline transcription, local LLM templates, and Markdown sync. Archive or delete old notes anytime to free up slots. • Pro Lifetime ($19.99 One-Time): Unlock unlimited notes and unlimited recording duration forever. No subscriptions, ever.