simple-auto-subtitle 0.1.0

Automatic subtitle generation using Gemini AI

Simple Auto Subtitle (Coauthor: Claude)

Automatic subtitle generation tool that extracts audio from video files and generates subtitles using Google's Gemini AI models.

Features

• Audio Extraction: Automatically extracts audio from video files using ffmpeg
• AI Transcription: Uses Gemini 2.5 models (Flash or Pro) for accurate speech-to-text with precise timestamps
• Multiple Output Formats:
  • .srt subtitle files
  • Videos with embedded subtitles (toggleable)
  • Videos with burnt-in subtitles (permanent)
  • Extracted audio files (.wav)
• Model Selection: Choose between Gemini Flash 2.5 (faster) or Pro 2.5 (more accurate)
• Structured Output: Uses Pydantic models for reliable JSON parsing
• Environment Variables: Supports .env files for API key management

Installation

From PyPI (Recommended)

pip install simple-auto-subtitle

From Source

git clone https://github.com/yunfanye/auto_subtitle.git
cd auto_subtitle
pip install -e .

Development Installation

git clone https://github.com/yunfanye/auto_subtitle.git
cd auto_subtitle
pip install -r requirements.txt
pip install -e .

Setup

1. Install ffmpeg (required for audio/video processing):

   • macOS: brew install ffmpeg
   • Ubuntu: sudo apt install ffmpeg
   • Windows: Download from ffmpeg.org

2. Set up Gemini API Key: Create a .env file in your working directory:

   GEMINI_API_KEY=your_api_key_here
   

Usage

Basic Usage

# Process default video (test.mp4) with Flash model
simple-auto-subtitle

# Process specific video file  
simple-auto-subtitle my_video.mp4

# Use Pro model for better accuracy
simple-auto-subtitle my_video.mp4 --model pro

# Alternative command
auto-subtitle my_video.mp4

Command Line Options

simple-auto-subtitle [video_file] [options]

Arguments:
  video_file              Video file to process (default: test.mp4)

Options:
  --api-key API_KEY       Gemini API key (or set GEMINI_API_KEY env var)
  --output, -o OUTPUT     Output SRT file path (default: video_name.srt)
  --model {flash,pro}     Gemini model: flash (faster) or pro (more accurate)

Output Files

For input video my_video.mp4, the tool generates:

• my_video.srt - Standard subtitle file
• my_video.wav - Extracted audio file
• my_video_embedded.mp4 - Video with embedded subtitles (can be toggled on/off)
• my_video_captioned.mp4 - Video with burnt-in subtitles (always visible)

Model Comparison

Model	Speed	Accuracy	Cost	Best For
Flash	Fast	Good	Lower	Quick processing, bulk videos
Pro	Slower	Better	Higher	High-quality transcription, important content

Requirements

• Python 3.8+
• ffmpeg
• Google Gemini API key
• Required Python packages: google-genai, pydantic, python-dotenv