Audio Transcription & Translation — User Guide

Overview

This app transcribes and optionally translates audio files. You configure your settings in the sidebar on the left, upload your files, run processing, then download the results.

Step 1: Configure Settings (Left Sidebar)

configuration

Model

Choose the Whisper model to use for transcription:

model
  • turbo (default) — use this for English-only audio when you want faster processing.
  • large-v3 — use this for non-English audio, or when you need the highest accuracy. Slower but more powerful.

Language

Select the language spoken in your audio files.

language
  • Auto-detect (default) — the model will guess the language automatically. Works well in most cases.
  • If you know the language, select it from the dropdown. This is more accurate than auto-detect and recommended when processing non-English audio.

Options

Speaker Recognition

speaker recognition
  • Check this box only if your audio has more than one speaker and you want the transcript to identify who said what (e.g. “SPEAKER_01: Hello”).
    • If there is only one speaker, leave this unchecked. Enabling it for single-speaker audio can reduce accuracy and may incorrectly split one person’s speech across multiple “speakers.”
  • When enabled, a Number of Speakers field appears:
    • Enter 0 to let the app auto-detect how many speakers there are.
    • Enter the exact number of speakers (e.g. 2) for more accurate results. Use this whenever you know how many people are in the recording.
    • Maximum supported: 20 speakers.

Translate to English

translation
  • Check this box if you want the output translated into English, regardless of the source language.
  • Can be combined with Speaker Recognition — the transcript will show speaker labels and English text together.

Step 2: Upload Your Audio Files

Supported formats: MP3, MP4, M4A, WAV, FLAC, OGG, WEBM

  1. In the File Upload section, click Upload.

    browse

  2. You can select multiple files at once.

The Files in ./data panel on the right shows everything currently queued for processing. From here you can:

view_files
  • Click Refresh to update the list.
  • Click X next to a file to remove it individually.
  • Click Delete All to clear every file in the queue.

Step 3: Run Transcription

Once your files are uploaded and your settings are configured, click the Start Transcription button.

start
  • A progress bar and status message will appear showing which file is being processed and how many remain.
  • Logs appear at the bottom of the page in real time.
  • Processing time depends on file length, the model chosen, and whether speaker recognition is enabled.

process
Note

Do not change settings or navigate away while processing is running.

Step 4: Download Results

When processing finishes, the Download Results section shows your output files.

  • Download All (ZIP) — downloads a transcriptions.zip archive containing your results:
    • .srt and .txt files for your output
    • if translation was used, the transcription in the original language as well as the translated English file
    • a reference.txt file with a research blurb and methodology

download_all
  • Click Delete All Outputs to remove all output files.

delete_all

Processing Logs

The Processing Logs section at the bottom records what happened during each run — which files were processed, any errors, and when processing completed.

logs

Click Clear Logs to reset the log view.

logs

Quick Reference

Goal Settings
Transcribe English audio, fast Model: turbo, Language: English
Transcribe non-English audio Model: large-v3, Language: select yours
Identify speakers in a recording Enable Speaker Recognition; set number of speakers if known
Get an English translation of foreign audio Enable Translate to English
Transcribed translation with speaker labels Enable both Speaker Recognition and Translate to English