🎬 VideoText Maker — Ultra‑Lightweight Text‑to‑Video Engine

Tech stack badges

---

✨ Overview

VideoText Maker is a blazing‑fast, low‑RAM text‑to‑video generator designed for cloud environments, CI pipelines, and lightweight VPS setups.
It produces narrated videos using edge‑tts, Pillow, and FFmpeg — without heavy frameworks like MoviePy.

This build is engineered for extreme efficiency:

• ⚡ ~30MB RAM usage (instead of ~500MB)
• 🚀 Zero in‑memory video buffers
• 🔥 Stream‑based pipeline
• 🧩 Perfect for Railway, Render, Fly.io, VPS, CI/CD

---

🚀 Key Features

• 🎙️ High‑quality TTS via edge‑tts (Microsoft neural voices)
• 🖼️ Text‑to‑frame rendering using Pillow
• 🎞️ Video assembly via FFmpeg (chunked, stream‑safe)
• 🧠 Low‑RAM architecture — no MoviePy, no numpy
• 🌐 Web UI included (Flask)
• 📦 Cloud‑ready (Procfile included)

---

🏗️ Architecture

A fully streaming pipeline:

Text → TTS → PNG frames → MP4 chunks → Final video

Components

• edge‑tts (CLI)
  Stable subprocess‑based TTS generation.

• Pillow
  Lightweight frame rendering without numpy overhead.

• FFmpeg
  Chunked MP4 assembly + final concat (no re‑encoding).

---

📂 Project Structure

.
├── app.py
├── requirements.txt
├── templates/
│   └── index.html
├── static/
├── output/
└── temp/

---

⚙️ Installation

🐧 Linux / WSL

sudo apt update
sudo apt install ffmpeg

python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

---

🪟 Windows

1. Install:

   • Python 3.10+
   • FFmpeg (add to PATH)

2. Verify:

ffmpeg -version

3. Install dependencies:

python -m venv venv
venv\Scripts\activate
pip install -r requirements.txt

---

🍎 macOS

brew install ffmpeg

python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

---

▶️ Run Locally

python app.py

Open in browser:

http://localhost:5000

---

🌐 Deployment (Railway / Heroku)

Procfile

web: python app.py

For production Flask:

web: gunicorn app:app

---

📉 Why Not MoviePy?

MoviePy:

• ❌ ~500MB RAM
• ❌ Slow rendering
• ❌ Unstable in containers

This project:

• ✅ ~30MB RAM
• ✅ FFmpeg streaming
• ✅ CI/CD‑friendly
• ✅ Zero heavy dependencies

---

🧠 Implementation Highlights

• Audio and video generation are fully separated
• No large in‑memory buffers
• Subprocess‑based FFmpeg pipeline
• Stream‑copy concatenation (no re‑encoding)
• Designed for low‑RAM cloud environments

---

🔧 Roadmap

• Parallel TTS generation
• Subtitle (SRT) support
• Improved Web UI
• Docker image

---

📜 License

MIT

---