Under the Hood

How SpokenAct Works

From voice to organized notes in seconds — here's exactly what happens under the hood.

Step 1

Record

One tap. That's it.

You tap the record button and start talking. SpokenAct captures your audio with a real-time waveform visualizer so you can see your voice in motion. No account setup, no cloud connection, no configuration. Just press and speak.

  • One-tap recording from the home screen
  • Real-time waveform visualization
  • Works offline — no internet required
  • No setup, no permissions fuss

Recording interface

Step 2

On-Device Transcription

Private. Fast. Free.

The moment you stop recording, Apple's SpeechAnalyzer framework converts your speech to text entirely on your iPhone. No audio is uploaded. No servers are involved. The transcription engine is the same one that powers Siri — it runs on the Neural Engine built into every modern iPhone.

  • Powered by Apple SpeechAnalyzer (on-device Neural Engine)
  • Audio never leaves your phone — true privacy
  • Works without an internet connection
  • Free and unlimited — no per-minute charges, ever

On-device processing

Step 3

AI Processing

Your transcript becomes structured intelligence

When you choose to summarize a note, only the transcript text — never the audio — is sent to our secure server. GPT-4o Mini analyzes the text and returns four things in seconds: a concise summary, extracted action items, key points, and topic tags. You must explicitly tap “Summarize” to trigger this — nothing is sent automatically.

  • Powered by GPT-4o Mini — fast, accurate, cost-efficient
  • Only transcript text is sent, never your audio recording
  • You control when AI runs — nothing happens without your tap
  • Returns: summary, action items, key points, and topic tags

What AI returns

Summary

Concise 2-3 sentence overview

Action Items

Extracted to-dos as a checklist

Key Points

Important ideas highlighted

Topic Tags

Auto-generated categories

AI-generated insights

Privacy & Security

We designed SpokenAct so your most personal data — your voice — stays on your device. Here is exactly what lives where.

Stays on your iPhone

  • Audio recordings

    Your voice never leaves your device. Recordings are stored locally and are never uploaded to any server.

  • Transcription processing

    Apple SpeechAnalyzer runs entirely on the Neural Engine in your iPhone. No cloud, no third parties.

  • Raw transcript text

    The transcript is stored on your device. It is only sent to our server if you explicitly tap Summarize.

Cloud (only when you choose)

  • Transcript text for AI

    When you tap Summarize, the text transcript is sent to our secure Supabase edge function, which calls GPT-4o Mini.

  • AI-generated results

    Summaries, action items, key points, and tags are stored in your encrypted Supabase account so they sync across sessions.

  • Never stored permanently on AI servers

    OpenAI does not retain your data for training. Transcripts are processed and discarded — not stored on their servers.

The bottom line

Your voice recording and transcription happen 100% on-device. AI summarization only runs when you ask for it, and only the text transcript is sent — never your audio. You are always in control.

Try it yourself

Record, transcribe, and get your first 3 AI summaries free. No credit card. No commitment.

Free forever to record & transcribe. AI summaries included with Premium.