Audio Recorder & Player

Record depositions, client calls, and meetings directly in SafeAppeals. Transcribe with local AI—no API costs, fully offline after first use.

Audio Recording

Record directly in the app with start, stop, and pause controls

Inline Playback

Play recordings with seek bar, time display, and volume control

AI Transcription

Local Whisper transcription—no cloud, no API costs


Opening the Audio Recorder

Access the Audio Recorder in three ways:

  • Keyboard: Press Ctrl+Shift+R
  • Command Palette: F1 → "Open Audio Recorder"
  • Activity Bar: Click the microphone icon in the left sidebar

Recording Audio

Starting a Recording

  1. Click the large Start Recording button
  2. A pulsing red indicator shows recording is active
  3. The timer displays elapsed time (MM:SS)

Recording Controls

Start

Begin capturing audio from your microphone

Pause / Resume

Temporarily pause and resume recording

Stop

End recording and save to your workspace

Recording Tips

Use a quality microphoneBetter audio quality leads to more accurate transcription results.
Minimize background noiseRecord in a quiet environment for cleaner audio.
Speak clearlyModerate pace and clear enunciation improve transcription accuracy.
Position microphone centrallyFor interviews, place the microphone between speakers.

Playing Recordings

Each recording card has integrated playback controls. No need to open a separate audio player.

Playback Controls

ControlDescription
Play / PauseStart or pause audio playback
Seek BarClick anywhere to jump to that position in the audio
Time DisplayShows current time / total duration
VolumeAdjust playback volume with slider

Tip: Audio loads lazily on first play to save memory. Blob URLs are automatically cleaned up when you close the panel.


Transcribing Audio

SafeAppeals uses Whisper distil-large-v3.5 for local transcription. No API costs, fully offline after the model downloads.

Starting Transcription

  1. Find the recording in your list
  2. Click the Transcribe button (speech bubble icon)
  3. Wait for processing (status shows "Transcribing...")

First-Time Setup

On your first transcription, the Whisper model (~1.5GB) downloads automatically:

DownloadingModel downloading (shows percentage)
LoadingLoading model into memory
ProcessingAnalyzing audio and generating transcript
CompleteTranscript ready to view and export

Whisper Model Details

PropertyValue
Modeldistil-whisper-large-v3.5-ONNX
Size~1.5GB (downloaded once)
Accuracy~7% Word Error Rate
Speed~1.5x faster than Whisper-Large-v3
Cost$0 — Fully local, no API calls

Importing Audio Files

Already have audio recordings? Import them to transcribe and manage alongside your new recordings.

Drag and Drop

  1. Open the Audio Recorder panel
  2. Drag audio files from your file explorer
  3. Drop them onto the recorder panel
  4. Files are imported and appear in your list

File Picker

  1. Click the Import Audio button
  2. Select one or more audio files
  3. Click Open to import

Supported Formats

.wavBest quality, no conversion
.mp3Common, widely compatible
.m4aApple/iOS recordings
.oggOpen format
.webmWeb recordings
.flacLossless compression

Exporting Transcripts

Export transcripts in multiple formats for different use cases.

FormatBest For
Word (.docx)Professional documents, legal filings, editing
Plain Text (.txt)Simple text, copy/paste, no formatting
Subtitles (.srt)Video editing, timestamped captions
JSON (.json)Integration, structured data with metadata

Recording Status

Each recording shows its current status:

PendingNot yet transcribed
TranscribingTranscription in progress
CompletedTranscript ready
FailedError during transcription

Per-Workspace Storage

Recordings are stored per workspace, keeping your case files organized and isolated:

📁 Your Case Folder/
├── 📄 Documents
├── 📄 Evidence
└── ...

📁 ~/.safe-appeals-navigator/databases/workspaces/{workspaceId}/
├── 📄 audio_recordings.db    ← Recording metadata
└── 📁 recordings/            ← Audio files
    ├── recording_001.wav
    ├── recording_002.wav
    └── ...
  • Separate for each case — Opening a different folder shows different recordings
  • Private — No cross-case access
  • Portable — Recordings stay with their case

AI Integration

The AI assistant can help manage your recordings. Try these prompts:

"Transcribe my latest recording"

Start transcription on the most recent audio

"Summarize the key points from the deposition recording"

Get AI summary after transcription

"Export the client call transcript as a Word document"

Generate DOCX export

"What did the doctor say about the injury in the medical recording?"

Query transcript content

Important: Always verify transcripts. AI transcription may contain errors, especially with technical terms, names, or accented speech.


Privacy & Offline

All audio processing happens locally on your machine:

No Cloud Uploads

Audio files never leave your computer. Recording and playback are completely local.

Offline Transcription

Whisper runs locally—no API calls, no transcription costs. Works without internet after model download.


Use Cases

Depositions

  • Record the deposition directly in SafeAppeals
  • Transcribe for searchable text
  • Export SRT for video sync if recorded on video
  • Export DOCX for official transcript

Client Calls

  • Record phone or video calls (check local consent laws)
  • Transcribe to capture key details
  • Link to case timeline events

Meetings & Interviews

  • Record witness interviews
  • Capture meeting notes automatically
  • Search transcripts later with AI chat

Troubleshooting

No audio captured

  • Check microphone permissions in system settings
  • Verify the correct microphone is selected
  • Ensure no other app is using the microphone

Transcription failed

  • Check if the recording has audio content
  • Verify sufficient disk space for model cache (~2GB)
  • Restart the app and try again

Slow transcription

  • First transcription loads the model (slower)
  • Subsequent transcriptions are faster
  • Longer recordings take proportionally longer

Import failed

  • Verify the file format is supported (WAV, MP3, M4A, OGG, WEBM, FLAC)
  • Check if the file is corrupted
  • Try converting to WAV format first

Keyboard Shortcuts

ActionShortcut
Open Audio RecorderCtrl+Shift+R

Next Steps