> ## Documentation Index
> Fetch the complete documentation index at: https://seedscroll.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Audio Transcription

> Record sermons, spoken notes, or audio files and transcribe them on-device to create study notes.

Seed Scroll can transcribe audio into text using Whisper, an on-device AI model. Transcription happens entirely on your device — your audio is never uploaded to a server.

## Record and transcribe

<Steps>
  <Step title="Open the Studies tab">
    Tap **Studies** in the bottom navigation bar.
  </Step>

  <Step title="Tap the microphone button">
    Tap the microphone icon in the floating action menu to start a new recording.
  </Step>

  <Step title="Record your audio">
    Speak clearly. Tap the stop button when you're done. Recordings can be up to 2 hours long.
  </Step>

  <Step title="Transcribe">
    After stopping, tap **Transcribe**. Whisper processes the audio on your device — this may take a minute or two for longer recordings.
  </Step>

  <Step title="Review and save">
    The transcript appears in the note editor. Edit it as needed, add hashtags and verse tags, then save.
  </Step>
</Steps>

## Import an audio file

You can also import an existing audio file instead of recording live.

<Steps>
  <Step title="Open Generate Study">
    From the **Studies** tab, tap **+** → **Generate Study**.
  </Step>

  <Step title="Select Import audio">
    Tap the audio import option and choose a file from your device.
  </Step>

  <Step title="Transcribe and generate">
    The app transcribes the audio and generates a structured study note from the transcript.
  </Step>
</Steps>

## Supported audio formats

MP3, MP4, WAV, AAC, M4A, FLAC, OGG

## Tips for accurate transcription

<Tip>
  Record in a quiet environment with the microphone close to the speaker. Background noise and multiple overlapping voices reduce accuracy.
</Tip>

<Tip>
  For long recordings (1+ hour), expect transcription to take several minutes on older devices. Keep the app in the foreground during transcription.
</Tip>

<Note>
  The Whisper model is downloaded to your device on first use. It requires approximately 75 MB of storage and a network connection for the initial download.
</Note>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.