The "sagar_builds" Subtitle Automation Guide

If you've seen my 50th reel, you know that retention is all about visual pace. One-word-at-a-time subtitles are the industry standard for keeping eyes glued to the screen, but doing it manually is a nightmare.

I built this tool to automate that grind using OpenAI's Whisper. This guide covers how to run it anywhere—including a full Linux environment inside your Android phone.

📺 Video Guide

🚀 The Logic

The script uses Whisper (Turbo) to handle transcription and generates a .srt file where every segment is exactly one word.

• Punctuation Stripping: Automatically removes commas for a cleaner look.
• Industry Standard CLI: Uses -i and -o flags for a professional workflow.

🛠️ Installation & Setup (PC/Mac)

1. Prerequisites

You must have FFmpeg installed (it handles audio extraction for the AI).

# Windows
choco install ffmpeg

# macOS
brew install ffmpeg

2. Install the Generator

git clone https://github.com/SAGAR-TAMANG/minimalist-one-word-subtitle-generator.git
cd minimalist-one-word-subtitle-generator
pip install -U openai-whisper

📱 The Termux "Ubuntu" Workflow (Android)

To avoid dependency errors on Android, we use PRoot Distro to run a real Ubuntu environment inside Termux. This is the most stable way to run AI on mobile.

Step 1: Install the Linux Environment

Paste these commands in Termux to set up your virtual Linux machine:

pkg update && pkg upgrade -y
pkg install proot-distro -y
proot-distro install ubuntu
termux-setup-storage

Step 2: Login & Setup Linux

Now, "log in" to your new Ubuntu system and install the necessary tools:

proot-distro login ubuntu

# Now inside Ubuntu:
apt update && apt upgrade -y
apt install python3 python3-pip ffmpeg git -y

Step 3: Clone the Tool

git clone https://github.com/SAGAR-TAMANG/minimalist-one-word-subtitle-generator.git
cd minimalist-one-word-subtitle-generator
pip3 install openai-whisper --break-system-packages

💻 Usage (CLI)

Run the script using these flags. In Ubuntu (Termux), your phone's files are bridged to the /sdcard path.

| Flag | Meaning | Example | | --- | --- | --- | | -i | Input Video | -i "/sdcard/Download/video.mp4" | | -o | Output Folder | -o "/sdcard/Download/" | | -m | AI Model | -m turbo (Fastest/Best) |

The Pro Command:

python3 main.py -i /sdcard/Download/raw_reel.mp4 -o /sdcard/Download/ -m turbo

🎨 Importing to Your Editor

1. VN Editor: Tap Text > SRT > Import. Apply a "Pop" animation to the entire track.
2. CapCut: Go to Text > Local Captions > Upload. Use "Batch Edit" to style every word simultaneously.

Builder Tip: I usually set the text to All Caps and use a bold font like The Bold Font or Komika Axis to maximize the impact.

🤝 Contributing

This is an open-source project. If you want to add auto-emoji placement or color-coding for keywords, open a PR on my GitHub!

Built by Sagar Tamang