Veo 3.1 Automated Video Pipeline

An automated video generation pipeline that converts ideas into videos using AI.

Features

• 📱 Mobile control via Telegram bot
• 🤖 AI-powered prompt generation (Gemini)
• 🎬 Video generation with Veo 3.1 API (15-key fallback)
• 💾 SQLite state persistence
• ⚡ Background processing (fire-and-forget)
• 🏥 Health check endpoint

Project Structure

veo_pipeline/
├── src/
│   └── main.rs           # Main application code
├── config/
│   └── api_keys.toml     # API keys configuration
├── videos/               # Downloaded videos (created at runtime)
├── Cargo.toml            # Rust dependencies
├── pipeline.db           # SQLite database (created at runtime)
└── README.md             # This file

Setup

1. Install Rust

curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

2. Configure API Keys

Edit config/api_keys.toml and add your keys:

• Veo 3.1 API keys (at least 1)
• Gemini API key for prompt generation

3. Set Telegram Bot Token

Get a token from @BotFather on Telegram.

Mac/Linux:

export TELOXIDE_TOKEN=your_telegram_bot_token

Windows (PowerShell):

$env:TELOXIDE_TOKEN="your_telegram_bot_token"

4. Update Veo Endpoint

In src/main.rs, update the veo_url variable with your actual Veo 3.1 endpoint.

5. Run the Project

cargo run

Usage

1. Open Telegram on your mobile
2. Send a message to your bot: /idea A futuristic city at sunset
3. The bot will reply instantly confirming your idea
4. Wait for the background processing to complete
5. The bot will notify you when the video is ready

Manual Testing Checklist

Phase 1-2: Server & Config

• Terminal prints Loaded X Veo API keys.
• Browser at http://localhost:3000/health shows {"status":"ok"}

Phase 3-4: Telegram & AI Prompt

• Bot replies instantly: ✅ Idea accepted...
• Terminal prints 📝 Prompt generated: ...

Phase 5-6: Veo Fallback & Download

• Terminal logs key fallback attempts
• Video file appears in videos/ folder
• Video plays correctly in media player

Phase 8: SQLite State

• Open pipeline.db with DB Browser for SQLite
• Check jobs table has correct data and status

Next Steps

• Phase 7: YouTube OAuth2 integration (coming soon)
• Phase 10: Cloud deployment

Tech Stack

• Language: Rust
• Web Framework: Axum
• Async Runtime: Tokio
• Bot Framework: Teloxide
• Database: SQLite (rusqlite)
• HTTP Client: reqwest
• AI: Google Gemini
• Video: Veo 3.1

License

MIT