alibaba/lumenx

alibaba/lumenx has 1,455 GitHub stars and 330 forks. GitHot measured +4 stars in 24 hours. Last checked 2026-11-06 08:14 UTC. Last push 2026-10-29.

How these counts are measured

LumenX Studio，AI 短漫剧一站式生产平台。它能够将小说文本转化为动态视频，打通了从剧本分析、角色定制、分镜绘制到视频合成的完整创作链路。

1,455 stars 330 forks +4 in 24h

Measured star growth

in 24h: +4
in 7 days: Update delayed
in 15 days: Update delayed
in 1 month: Update delayed

Net star changes over overlapping periods, not daily totals. Missing measurements are left blank. Stars measure attention, not software quality.

README_EN.md

Original

LumenX

AI-Native Motion Comic & Video Creation Platform

Render Noise into Narrative

English · 中文 · Changelog · Contributing

LumenX is an AI-native motion comic & video creation platform. It transforms creative text into publishable dynamic videos, providing a complete workflow from script analysis to final export, while also supporting standalone image/video generation.

LumenX currently includes two core modules:

Module	Purpose
LumenX Studio	Pipeline-first comic/video production (Script → Storyboard → Assets → Video → Export)
LumenX Playground	Standalone image/video generation workbench (no project context required)

✨ Core Capabilities

🎬 Studio — Full Pipeline Production

Deep Script Analysis — LLM auto-extracts characters/scenes/props, generates structured storyboards
Art Direction Control — Custom visual styles with global consistency
Multi-model Asset Generation — Character turnarounds, scene establishing shots, prop references
AI Video Generation — I2V / R2V multi-mode video generation + batch candidates
Smart Dubbing — CosyVoice / Qwen3-TTS multi-voice dialogue synthesis
One-click Export — Timeline editing + FFmpeg merging

🎨 Playground — Standalone Generation Workbench

6 Generation Modes — Image, Text-to-Video, Image-to-Video, Reference-to-Video, Video Editing
10+ AI Models — GPT-Image-2, Wan 2.7, Seedance 2.0, Kling V3, Vidu Q3, HappyHorse, etc.
Dynamic Parameters — Per-model parameter configuration (size/resolution/duration/quality)
Concurrent Tasks — Multiple tasks execute simultaneously with real-time status tracking
Prompt Templates — Save/reuse/favorite/history
Gallery View — Grid/gallery toggle + detail panel

🎨 v1.2.1 Visual Identity Refresh

Before	After

Neon gradient lotus · Soft curves	Cyber Brutalism · Angular geometry · Circuit textures

📸 Screenshots

Studio Storyboard	Playground

🎯 Supported AI Models

Provider	Models	Capabilities
DashScope	Wan 2.7 Image/Video, Qwen Image 2.0, HappyHorse 1.0	T2I, I2I, I2V, R2V, T2V, V2V
DashScope	Kling V3	I2V, R2V
DashScope	Vidu Q3 Pro / Turbo	I2V, R2V
DashScope	PixVerse V6 / C1	I2V, R2V
MuleRun	Seedance 2.0	T2V, I2V, R2V
MuleRun	GPT-Image-2	T2I, I2I (up to 4K)
Kling Direct	Kling V3	I2V, R2V
Vidu Direct	Vidu Q3 Pro / Turbo	I2V, R2V
DashScope	CosyVoice, Qwen3-TTS	TTS Dubbing
DashScope	Qwen 3.7 Plus	Script Analysis, Prompt Polish

🚀 Quick Start

Prerequisites

Python 3.11+
Node.js 18+
FFmpeg (for video processing)

One-command Launch

# Clone
git clone https://github.com/alibaba/lumenx.git
cd lumenx

# Configure API Key
cp .env.example .env
# Edit .env, fill in DASHSCOPE_API_KEY (required)

# Start (backend on 17177 + frontend on 3008, auto-opens browser)
npm run dev

Or start separately:

# Backend
pip install -r requirements.txt
./start_backend.sh  # http://localhost:17177

# Frontend
cd frontend && npm install && npm run dev  # http://localhost:3008

Access

Studio: http://localhost:3008
Playground: http://localhost:3008/#/playground
API Docs: http://localhost:17177/docs

⚙️ Configuration Modes

LumenX uses a local-first architecture. The minimal setup requires only one API key.

Mode	Required	Available Capabilities
Basic	`DASHSCOPE_API_KEY`	Wan/Qwen/HappyHorse/PixVerse/Kling(proxy)/Vidu(proxy) + TTS
+ MuleRun	+ `mulerun login` or `MULEROUTER_API_KEY`	+ Seedance 2.0 + GPT-Image-2
+ Kling Direct	+ `KLING_ACCESS_KEY` + `KLING_SECRET_KEY`	Kling direct connection
+ Vidu Direct	+ `VIDU_API_KEY`	Vidu direct connection
+ OSS	+ Alibaba Cloud OSS credentials	Cloud media mirror + signed URLs

Detailed Configuration

All settings can be configured via:

Development: .env file in project root
In-app Settings: Settings page (saves to ~/.lumen-x/config.json)

MuleRun supports two authentication methods:

CLI mode (recommended): npm i -g @mulerunai/cli && mulerun login
API Key mode: Enter muk-... format key in Settings page

🏗️ Architecture

Directory Structure

lumenx/
├── frontend/                  # Next.js Frontend
│   └── src/components/
│       ├── modules/playground/   # Playground module
│       ├── modules/              # Studio business modules
│       └── layout/               # Global layout
├── src/
│   ├── apps/comic_gen/        # Studio backend (API + Pipeline)
│   ├── apps/playground/       # Playground backend (API + Service)
│   ├── models/                # AI model adapters (Wanx/Kling/Vidu/MuleRouter)
│   └── audio/                 # TTS voice synthesis
├── config/model_catalog/      # Model catalog (YAML → JSON)
└── output/                    # Generated outputs (local storage)

📖 Documentation

Document	Description
User Manual	Feature usage guide
API Docs	Swagger UI
Model Onboarding	New model integration guide
Catalog Architecture	Model catalog design
Playground PRD	Playground design document

🤝 Contributing

Contributions are welcome! Please read our Contributing Guide.

Bug Reports: GitHub Issues
Feature Requests: GitHub Discussions
Email: [email protected]

📄 License

MIT License

Made with ❤️ by StarLotus · Alibaba Group