ATH-MaaS/Pixelle-Video

🚀 AI 全自动短视频引擎 | AI Fully Automated Short Video Engine

What it solves

Pixelle-Video is an automated short-video engine that allows users to create complete videos from a single topic or keyword. It removes the need for manual scriptwriting, image generation, voiceover synthesis, and video editing, making professional-looking short-form content accessible to those with zero editing experience.

How it works

The project uses a modular pipeline to transform text into video through four main stages:

  1. Script Generation: An LLM (such as GPT, DeepSeek, or Qwen) creates a video script based on the user's topic.
  2. Visual Planning: The system plans the necessary imagery for each segment of the script.
  3. Frame Processing: AI models generate the visual assets. This can be done via ComfyUI workflows, RunningHub, or direct API calls to providers like OpenAI, DashScope, or Kling AI for images and videos.
  4. Video Synthesis: The system combines the generated visuals, synthesized speech (via TTS engines like Edge-TTS or Index-TTS), and background music into a final video using HTML-based templates for layout and style.

Who it’s for

  • Content Creators: People who want to produce short-form videos quickly without manual editing.
  • AI Enthusiasts: Users who want to integrate various AI models (LLMs, Image/Video Gen, TTS) into a single automated workflow.
  • Developers: Those who want to customize video generation pipelines using ComfyUI workflows or HTML templates.

Highlights

  • Fully Automated: Generates scripts, images, videos, and audio from a single prompt.
  • Flexible Model Integration: Supports a wide range of LLMs (Ollama, GPT, DeepSeek) and media models (Wan 2.1, Kling, DashScope).
  • Modular Architecture: Allows users to swap out TTS engines, image generators, and video templates.
  • Multi-modal Capabilities: Includes advanced modules for digital human talking heads, image-to-video generation, and motion transfer.
  • Customizable Styles: Offers various HTML templates for different video orientations (vertical, horizontal, square) and visual styles.

Related

  • Project
    xhongc/ai_story

    An AI-driven automation platform that transforms a story theme into a complete video by automating scriptwriting, storyboarding, image generation, and video production.

  • Project
    Agents365-ai/video-podcast-maker

    An automated pipeline that transforms a topic into a professional 4K video podcast by coordinating AI research, script generation, TTS audio, and Remotion video rendering.

  • Project
    myccarl/ai-shortVideo-pipeline

    An end-to-end automated pipeline for producing short-form AI videos, covering everything from topic discovery and scripting to visual generation and post-production.

  • Project
    SamurAIGPT/Text-To-Video-AI

    An AI-powered video creation tool that transforms text prompts into fully edited videos with scripts, voiceovers, B-roll, and synchronized captions.

  • Project
    leamsigc/ShortsGenerator

    An automated pipeline for creating short-form videos that handles AI script generation, stock footage sourcing, TTS voiceovers, and social media scheduling.