Video Text Overlay App

A simple Streamlit application that allows you to:

1. Upload a video
2. Add text overlay (manually or generate using LLMs)
3. Convert text to speech
4. Combine everything into a final video

Features

• Video upload support (MP4, MOV, AVI)
• Text overlay with customizable appearance
• Text-to-speech conversion
• Integration with local LLMs (Ollama and LM Studio)
• Easy download of processed videos

Setup

Local Setup

1. Install dependencies:

pip install -r requirements.txt

2. Run the app:

streamlit run app.py

Docker Setup

1. Build the container:

docker build -t video-text-overlay .

2. Run the container:

docker run -p 8501:8501 video-text-overlay

LLM Integration

The app supports two local LLM options:

Ollama

• Install Ollama from: https://ollama.ai/
• Run Ollama locally
• Default endpoint: http://localhost:11434

LM Studio

• Install LM Studio from: https://lmstudio.ai/
• Run your chosen model locally
• Default endpoint: http://localhost:1234/v1/completions

Usage

1. Open the app in your browser (default: http://localhost:8501)
2. Choose your LLM settings in the sidebar (optional)
3. Upload a video file
4. Either:
   • Enter text directly in the text area
   • Or use an LLM to generate text by entering a prompt
5. Click "Process Video" to create the final video
6. Download the processed video using the download button

Notes

• The app creates temporary files that are cleaned up after processing
• Text overlay is centered and sized to 80% of video width
• Audio is generated using pyttsx3 (offline TTS)