Hi HN,

I built a voice-to-text tool that runs entirely in your browser. No account required for the free tier, no data sent to my servers.

Try it: https://voicetotextonline.com

Why I built this:

- Existing tools require signups, have minute limits, or cost money - Google Docs voice typing requires a Google account - Dragon costs $150-500 - Otter.ai has free tier limits

(A) Free Features (no account required):

1/ Core Transcription:

- Real-time voice-to-text using Web Speech API - 55+ languages supported - Auto-punctuation & sentence case options - Works offline after first load (PWA)

2/ AI Enhance (added based on user survey – 80% voted yes):

- Auto-fix grammar, punctuation & formatting - One-click cleanup of transcripts

3/ My Projects (local storage):

- Save transcripts to browser localStorage - Organize with folders (Notes, Work, Personal, etc.) - Custom folders & tags - Search across all transcripts - Edit, copy, download as TXT - 100% private – never leaves your device

- Export:

- Copy to clipboard - Download as TXT or DOCX

(B) Pro Features ($10/month or $1/hour pay-per-use):

1/ File Upload & Transcription:

- Upload audio/video files (MP3, WAV, M4A, MP4, MOV, AVI, MKV) - Up to 500MB per file - Batch upload (10 files at once) - Powered by AssemblyAI (95%+ accuracy) - 150 hours/month transcription

2/ Advanced Features:

- Real-time progress with ETA - Speaker labels - In-browser audio recording (5 min with pause/resume) - Translation to 25+ languages (GPT-4o)

3/ Export Formats:

- TXT, SRT, VTT, JSON with timestamps - Segment-level timestamp precision

4/ Cloud Storage:

- Transcription history in the cloud - 10 GB storage, 1,000 files/month

(C) Data & Privacy:

Free tier:

- All transcripts stored in browser localStorage only - Never touches our servers - 100% private

Pro tier:

- Audio files stored in Supabase (encrypted) - Files retained for 30 days for re-download, then auto-deleted - Transcripts stored permanently in your account - You can delete any transcript or your entire account anytime - We don't use your data for training

Tech stack:

- Next.js 14 (App Router) - Web Speech API (free real-time transcription) - AssemblyAI (Pro file transcription, 95%+ accuracy) - OpenAI GPT-4o (AI Enhance & translation) - Supabase (auth & storage) - Stripe (payments) - Tailwind CSS - Hosted on Vercel

Limitations:

- Real-time transcription doesn't work in Firefox (Web Speech API not supported) - Free tier accuracy depends on Chrome's speech engine

Would love feedback on UX, pricing, or feature ideas. Considering open-sourcing the core transcription component.