Gemini YouTube Automation

The project includes a GitHub Actions workflow that runs daily at 7:00 AM UTC. It:

• Generates lesson scripts using Gemini.
• Produces long-form and short YouTube videos.
• Uploads them automatically with appropriate thumbnails and metadata.

Project Structure

gemini-youtube-automation/
├── .github/
│   └── workflows/
│       └── main.yml         # GitHub Actions workflow configuration
├── src/                     # Source directory for Python modules
│   ├── init.py          # Initializes the 'src' package
│   ├── generator.py         # Code for generating content and video
│   └── uploader.py          # Code for uploading to YouTube
├── .gitignore               # Files and directories to ignore in version control
├── content_plan.json        # Contains topics for moving forward.
├── main.py                  # Main entry point to run the application
└── requirements.txt         # List of Python packages needed

Setup Instructions

1. Clone the repository and open a terminal in the project root (the folder that contains main.py).

2. Install dependencies (Python 3 recommended):

   pip install -r requirements.txt

3. Environment variables

   • GOOGLE_API_KEY (required): Google AI / Gemini API key.
   • PEXELS_API_KEY (optional): If unset, slide backgrounds fall back to solid colors.

4. YouTube upload (local)

   • Create a YouTube Data API OAuth Desktop app client in Google Cloud and download the JSON as client_secrets.json in the project root.
   • On the first run, the app opens a browser to complete OAuth and writes credentials.json for later runs.

Execution methods

Run locally

From the project root, with GOOGLE_API_KEY set (and optionally PEXELS_API_KEY):

python main.py

On Windows, if python is not on your PATH, try py main.py.

This loads or creates content_plan.json, produces up to one pending lesson (long + short video), uploads to YouTube, updates lesson status, and writes artifacts under output/. Ensure FFmpeg is available if MoviePy or audio conversion fails on your system.

Run on GitHub Actions (scheduled)

The repo is intended to run on a daily schedule (for example 7:00 AM UTC) via a workflow under .github/workflows/ (for example main.yml). Typical setup:

1. Add repository secrets for GOOGLE_API_KEY and, if you use stock imagery, PEXELS_API_KEY.
2. Provide YouTube credentials in CI the same way your workflow expects (for example writing client_secrets.json and/or credentials.json from secrets before python main.py), so src/uploader.py can refresh tokens non-interactively.
3. Push to the branch your workflow uses (usually main); use Actions → Run workflow if the workflow defines workflow_dispatch for manual runs.

Adjust steps to match your actual workflow file once it is committed.