oneword-ai 0.2.8

Create viral-style one-word subtitles for Reels, Shorts & TikToks using AI. Perfect for content creators!

🎬 OneWord AI - Production Subtitle Generator

Enterprise-grade visual subtitle generator for Local & VPS Environments

Generate viral-style subtitles locally or on your own server. No cloud notebooks, no timeouts.

Installation • Web UI • CLI • VPS Deployment

---

🚀 Installation

Local Setup (Windows/Mac/Linux)

1. Install Python 3.9+ and FFmpeg.

2. Install the package:

   pip install oneword-ai
   

   pip install oneword-ai

📱 ARM / Android Support (Termux)

1. Install Dependencies:

   pkg install python git clang make
   

2. Install ARM Package (Lightweight, No PyTorch):

   pip install oneword-ai-arm
   

3. Setup Whisper.cpp (One-Time):

   onewordai-setup-arm
   

---

📱 Web Interface (Mobile)

To run the lightweight Web UI on your phone:

onewordai-web-arm

Open http://localhost:8000 in your mobile browser.

---

🖥️ Web Interface (Desktop)

Start the local web server to use the visual interface:

oneword-web
# OR
python -m onewordai.api.main

Open http://localhost:8000 in your browser.

Features:

• ⚡ Local Processing: No file size limits, 100% privacy.
• 🎯 3 Modes: One Word, Two Word, Phrase.
• 🌍 Multi-Language: Supports English, Hindi, Urdu, Spanish.
• 🇮🇳 Special Hindi Model: Uses Oriserve/Whisper-Hindi2Hinglish-Prime.

---

⌨️ Command Line (CLI)

Perfect for batch processing or automation scripts:

# Process a single video
oneword-cli -i input_video.mp4

# Customize settings
oneword-cli -i input.mp4 -m medium -lang en --mode twoword

# Full help
# Full help
oneword-cli --help

ARM Devices (Termux)

Use the optimized ARM command:

# Process video
onewordai-process-arm -i video.mp4

# With settings
onewordai-process-arm -i video.mp4 -m base -lang hi

# Download other models
cd onewordai/arm/whisper.cpp
bash ./models/download-ggml-model.sh small

# Run Web UI
onewordai-web-arm

---

☁️ VPS / Server Deployment

Method 1: Docker (Recommended)

1. Build the image:

   docker build -t oneword-ai .
   

2. Run container:

   docker run -d -p 8000:8000 oneword-ai
   

Method 2: Systemd (Linux Servers)

Run as a background service on Ubuntu/Debian:

1. Create service file: sudo nano /etc/systemd/system/oneword.service

   [Unit]
   Description=OneWord AI Web Server
   After=network.target
   
   [Service]
   User=root
   WorkingDirectory=/root/oneword-ai
   ExecStart=/usr/local/bin/oneword-web
   Restart=always
   
   [Install]
   WantedBy=multi-user.target
   

2. Start service:

   sudo systemctl enable oneword
   sudo systemctl start oneword
   

---

📦 Models

Models are downloaded to ~/.cache/huggingface (or generic cache) on the first run.

Model	Size	Best For
medium	~1.5GB	General Purpose (Default)
large	~3GB	High Accuracy
Hindi2Hinglish	~1.5GB	Hindi to Hinglish Transcription

---

📜 License

MIT License. Free for commercial use.