CLI Reference¶

WhisperJAV provides several command-line tools.

whisperjav (Main CLI)¶

Transcribe video/audio files to SRT subtitles.

whisperjav [OPTIONS] INPUT [INPUT...]

Basic Options¶

Option	Default	Description
`INPUT`	(required)	One or more video/audio file paths
`--mode`	`balanced`	Pipeline: `balanced`, `fast`, `faster`, `fidelity`, `transformers`
`--sensitivity`	`aggressive`	Detection sensitivity: `aggressive`, `balanced`, `conservative`
`--language`	`ja`	Source language code (`ja`, `ko`, `zh`, `en`)
`--output-dir`	Same as source	Directory for output SRT files
`--output-format`	`srt`	Output format: `srt`, `vtt`, `both`

Model Options¶

Option	Default	Description
`--model`	(auto)	Whisper model: `large-v2`, `large-v3`, `turbo`
`--translate`	Off	Translate to English during transcription

Ensemble Options¶

Option	Default	Description
`--ensemble`	Off	Enable ensemble mode
`--ensemble-serial`	Off	Complete each file fully before starting the next
`--pass1-pipeline`	`balanced`	Pipeline for first pass
`--pass2-pipeline`	`qwen`	Pipeline for second pass
`--merge-strategy`	`smart`	Merge strategy: `pass1_primary`, `smart`, `full`, `pass2_primary`, `longest`

Processing Options¶

Option	Default	Description
`--cpu-only`	Off	Force CPU mode (no GPU)
`--async`	Off	Enable async processing
`--temp-dir`	System temp	Custom directory for temporary files
`--keep-temp`	Off	Keep intermediate files
`--debug`	Off	Enable debug logging

Examples¶

# Basic transcription
whisperjav video.mp4

# Fast mode with conservative sensitivity
whisperjav video.mp4 --mode faster --sensitivity conservative

# Ensemble with serial mode
whisperjav *.mp4 --ensemble --ensemble-serial --merge-strategy smart

# VTT output
whisperjav video.mp4 --output-format vtt

# Custom output directory
whisperjav video.mp4 --output-dir ./subtitles/

# CPU-only mode
whisperjav video.mp4 --mode faster --cpu-only

whisperjav-gui¶

Launch the GUI application.

whisperjav-gui

No additional arguments. All configuration is done through the GUI interface.

whisperjav-translate¶

Standalone subtitle translation tool.

whisperjav-translate [OPTIONS] -i INPUT.srt

Option	Default	Description
`-i`, `--input`	(required)	Input SRT file(s)
`--provider`	`deepseek`	AI provider
`--model`	(auto)	Model name
`--api-key`	(env var)	API key (or set via environment variable)
`--target-language`	`English`	Target language
`--tone`	`standard`	Translation tone: `standard`, `adult`

whisperjav-upgrade¶

Upgrade WhisperJAV to the latest version.

whisperjav-upgrade [OPTIONS]

Option	Description
`--check`	Check for updates without upgrading
`--wheel-only`	Code-only upgrade (skip dependency reinstall)
`--list-snapshots`	Show available rollback points
`--rollback`	Rollback to previous version
`--extras`	Upgrade specific extras only (e.g., `cli,gui`)

See Upgrade Guide for details.