Scriberr

A self-hosted audio transcription tool powered by WhisperX. Upload audio files through the web interface and get accurate transcriptions with speaker diarization support.

What You Can Do After Deployment

Visit your domain to access the transcription interface. Upload audio files and Scriberr will process them automatically.

Upload and transcribe: Drag and drop audio files (MP3, WAV, M4A, and more) into the web interface. Transcription starts automatically.

Speaker diarization: Identify different speakers in the audio. Each speaker's segments are labeled separately in the output.

Multiple languages: WhisperX supports transcription in many languages. The model auto-detects the spoken language or you can specify it manually.

Edit transcriptions: Review and edit transcription results directly in the web interface. Export to various formats.

SQLite storage: All data is stored locally in SQLite — no external database needed. Audio files and transcriptions persist in the data volume.

Configuration

• Data directory: Audio files and transcriptions are stored in the /app/data volume
• WhisperX environment: Model files are cached in the /app/whisperx-env volume (large initial download)
• Resource note: Transcription is CPU-intensive. Allow extra time for processing on smaller instances

License

MIT — GitHub