Pixelle-Video β€” AI Fully Automated Short Video Engine

Pixelle-Video β€” AI Fully Automated Short Video Engine

Pixelle-Video is an open-source platform that automates the entire short video creation pipeline β€” from topic input to finished video. No editing experience needed. Input a topic, and it handles scriptwriting, visual planning, image/video generation, TTS narration, and final composition automatically. With ~6.9k GitHub stars and growing, it’s one of the most popular AI video automation tools.

Source: GitHub - AIDC-AI/Pixelle-Video

How the Pipeline Works

Topic Input
    ↓
Script Generation (GPT / Qwen / DeepSeek / Ollama)
    ↓
Visual Planning (match visuals to each narrative segment)
    ↓
Image/Video Creation (AI-generated assets)
    ↓
Audio Synthesis (multi-language TTS)
    ↓
Composition (assembly + background music + styling)
    ↓
Final Video Output (vertical / horizontal / custom aspect)

Key Features

Feature Description
Zero editing No scripts or editing skills needed
Multi-model GPT, Qwen, DeepSeek, Ollama integration
Flexible output Vertical, horizontal, custom aspect ratios
Templates Pre-built templates for different content categories
Multi-language TTS Multiple voice options across languages
Custom materials Upload your own photos/videos for AI analysis
Digital human AI presenter narration with multilingual voices
Motion transfer Dynamic content from static images
Image-to-video Convert images to video with AI animation

Architecture

Built on a modular, composable design using ComfyUI-style workflows. Components can be swapped independently β€” replace image models with FLUX, TTS with ChatTTS, etc.

Installation

# Windows: Download integrated package from releases
# Execute start.bat β†’ web UI at localhost:8501

# Other platforms:
# Requires Python 3.10+, FFmpeg, UV package manager
git clone https://github.com/AIDC-AI/Pixelle-Video.git
cd Pixelle-Video
# Follow README for setup

How LearnAI Team Could Use This

  • Content creation β€” generate educational short videos from topic descriptions
  • Douyin/social media β€” automate the video pipeline for publishing
  • Teaching demos β€” quickly produce visual explanations of CS concepts
  • Compare with existing pipeline β€” benchmark against the manual Douyin workflow we’ve built

Real-World Use Cases

  • Content creators β€” produce daily short-form content without manual editing
  • Marketing teams β€” generate product explainer videos at scale
  • Educators β€” turn lecture topics into bite-sized video lessons
  • Social media managers β€” maintain consistent posting cadence with AI-generated content