Skip to content

Common Workflows

Step-by-step recipes for the most common tasks.

Quick Transcription (Simplest)

The fastest path from video to subtitles.

  1. Launch WhisperJAV GUI
  2. Drag a video file onto the app
  3. Leave all defaults (Balanced mode, Aggressive sensitivity, Japanese)
  4. Click Start
  5. SRT file appears next to your video

Time: ~2-5 minutes for a 2-hour video with a modern GPU.

High-Quality Transcription (Ensemble)

Uses two different ASR engines and merges their output for best accuracy.

  1. Go to the Ensemble tab
  2. Pass 1: Balanced pipeline (default)
  3. Enable Pass 2: Check the Pass 2 checkbox
  4. Pass 2: Select Qwen3-ASR
  5. Merge Strategy: Smart Merge
  6. Click Start

Both passes run, then results are intelligently combined. Takes roughly 2x the time of a single pass.

ChronosJAV Pipeline (Anime/JAV Content)

Dedicated pipeline using models trained on anime and JAV dialogue.

  1. Go to the Ensemble tab
  2. Pass 1 Pipeline: Select ChronosJAV
  3. Model: Choose from:
    • anime-whisper (~4GB) — best quality for anime/JAV
    • Kotoba v2.1 (~2GB) — lighter, with punctuation
    • Kotoba v2.0 (~2GB) — lighter, no punctuation
  4. Click Start

Tip

For maximum quality, use anime-whisper in Pass 1 and Qwen3-ASR in Pass 2 with Smart Merge.

Transcribe + Translate in One Step

Get translated subtitles without a separate step.

  1. Configure your transcription (any tab)
  2. On the Ensemble tab, check "AI-translate"
  3. Select your translation provider (DeepSeek, Gemini, etc.)
  4. Enter your API key if needed
  5. Click Start

Transcription runs first, then translation happens automatically on the result.

Translate an Existing SRT File

Use Tab 4 to translate subtitles you already have.

  1. Go to the AI SRT Translate tab (Tab 4)
  2. Click Add File(s) and select your .srt file
  3. Select a Provider and Model
  4. Enter your API key and click Test Connection
  5. Set Target Language (e.g., English)
  6. Choose Tone: Standard or Adult-Explicit
  7. Click Start

Note

The Adult-Explicit tone uses specialized instructions tuned for JAV dialogue with appropriate vocabulary.

Batch Processing (Multiple Files)

Process an entire folder of videos at once.

  1. Click Add Folder and select a folder containing videos
  2. All media files are added to the list
  3. Configure your pipeline settings
  4. Click Start

Files are processed sequentially. Each output SRT is saved next to its source video (or to your chosen output directory).

Serial Ensemble Mode

In Ensemble mode, enable Serial mode to complete each file fully (Pass 1 → Pass 2 → Merge) before starting the next. This lets you see results as they finish instead of waiting for the entire batch.

CPU-Only Mode (No GPU)

WhisperJAV works without a GPU, just slower.

  1. Go to the Advanced tab
  2. Check "Accept CPU-only mode"
  3. Use Faster mode for the best speed without GPU
  4. Click Start

Warning

CPU mode is 5-10x slower than GPU mode. A 2-hour video may take 30-60 minutes.

WebVTT Output

Generate VTT subtitles for HTML5 video players.

  1. Go to the Advanced tab
  2. Set Output Format to VTT or Both (SRT + VTT)
  3. Run your transcription as normal

The .vtt file is saved alongside the .srt file.