AI Video Generator

Version: 3.3 | Last Updated: 2025-12-06

An automated video creation tool that transforms a simple topic into a complete, share-ready video with an AI-generated script, professional voiceover, relevant visuals, and an eye-catching thumbnail.

Overview

The AI Video Generator transforms video content creation from a multi-hour, multi-tool process into a streamlined 20-minute workflow. It's designed for content creators who value production speed and quality but lack the time, budget, or technical skills for traditional video production.

Our key value proposition is a local-first, privacy-focused architecture. Unlike cloud-dependent solutions, our system runs primarily on your own hardware using free and open-source (FOSS) technologies like Ollama and local TTS. For those seeking enhanced quality, the system seamlessly integrates with powerful cloud services like Google Gemini and ElevenLabs, maintaining a hybrid "FOSS-first, cloud-enhanced" approach with zero mandatory subscription costs.

Core Features (Complete)

• Conversational AI Agent: Brainstorm and refine video ideas through a natural chat interface.
• Automated Script Generation: Get a professional, human-sounding script divided into logical scenes.
• LLM Configuration & Script Personas: Choose between local (Ollama) or cloud (Gemini) models and select from personas like 'Scientific Analyst' or 'Blackpill Realist' to define your content's tone.
• Voice Selection & Synthesis: Select from a diverse catalog of local (FOSS) or cloud (ElevenLabs) voices to narrate your script.
• AI-Powered Visual Sourcing: Automatically finds relevant B-roll from YouTube, using Google Cloud Vision to filter out talking heads, captions, and irrelevant content.
• Visual Curation UI: Review and select the perfect video clip for each scene from AI-powered suggestions.
• Automated Video Assembly: Automatically combines your selected visuals and voiceover into a final MP4 video using FFmpeg.
• Automated Thumbnail Generation: Instantly get a compelling thumbnail with a relevant background and your video's title.

Enhancement Features (In Development)

• Automated Background Music: Automatically selects, mixes, and applies topic-appropriate background music.
• AI-Generated SEO Toolkit: A VidIQ-style command center that generates optimized titles, descriptions, and tags, plus keyword research and pre-upload SEO audits.
• Automate Mode: A full, one-click automation pipeline from topic confirmation to final export.
• ElevenLabs TTS Integration: Use premium cloud-based voices from ElevenLabs as an alternative to the local TTS engine.
• Unified API Usage Dashboard: A single dashboard to monitor your usage and quotas for all integrated APIs (Gemini, YouTube, ElevenLabs).

Technology Stack

Frontend & Framework

• Next.js 15 - React framework with App Router
• TypeScript - Type-safe development
• Tailwind CSS v4 - Utility-first styling
• shadcn/ui - Accessible component library

AI & Data Processing

• LLM: Ollama (Llama 3.2) for local processing, Google Gemini for cloud-enhanced generation.
• TTS: KokoroTTS (local, FOSS), ElevenLabs (cloud).
• Visuals: YouTube Data API for sourcing, Google Cloud Vision API for filtering.
• Audio/Video: FFmpeg for assembly, yt-dlp for downloading.
• Databases: SQLite for project data, ChromaDB/LanceDB for future RAG capabilities.
• State Management: Zustand for lightweight client state.

Prerequisites

Required

• Node.js 18+
• Python 3.9+
• UV Package Manager
• Ollama (with a pulled model, e.g., llama3.2)
• FFmpeg 7.1+

Setup

# Clone repository
git clone https://github.com/AIfriendly/AIvideogen.git
cd AIvideogen/ai-video-generator

# Install Node.js dependencies
npm install

# Install Python dependencies (from parent directory)
cd ..
uv pip install -r requirements.txt
cd ai-video-generator

# Start development server
npm run dev

Open http://localhost:3000 to use the application.

Future Enhancements

We are actively working on expanding our capabilities with features like:

• Domain-Specific Video Sources: Adding official sources like DVIDS for military footage.
• Local Computer Vision: A FOSS alternative to Google Vision using MediaPipe and Tesseract.js for zero-cost, private analysis.
• RAG-Powered Channel Intelligence: A VidIQ-style system to analyze competitors, monitor trends, and generate scripts informed by your specific niche.

License

MIT License - See LICENSE file for details.