A modular, node-based content generation engine designed to transform creative ideas into high-quality visual content including stickers, memes, and personalized greetings. Powered by Alibaba Cloud AI technologies for an intuitive visual workflow experience.
This application leverages Alibaba Cloud Qwen - an advanced language model by Alibaba Cloud - as the core AI engine for text generation, prompt enhancement, and intelligent content reasoning. Combined with Alibaba's image and video generation capabilities, ADIS AI delivers a comprehensive solution for creating culturally-relevant Indonesian content.
ADIS AI provides a visual pipeline editor that enables users to create complex content generation workflows through an intuitive drag-and-drop interface. The application leverages Alibaba Cloud AI services (Qwen for text/reasoning, Wan for image/video generation) to power the content generation capabilities.
- Visual Node-Based Editor: Create and manage content generation pipelines using an intuitive canvas interface
- Typed Port System: Connection validation ensures data compatibility between nodes with distinct port types (text, prompt, image, video, style, media)
- Pipeline Execution Engine: Client-orchestrated execution with server-side AI processing, featuring topology sorting and state management
- Template Presets: Quick-start workflows for common use cases including Ramadan Wishes, Holiday Memes, AI Pet, and Custom Avatar generation
- Real-time Execution Feedback: Live progress updates and logging during pipeline execution
- Indonesian Cultural Context: Specialized templates and content generation for Indonesian cultural events and celebrations
The pipeline editor supports five distinct node categories:
| Category | Description | Color Code |
|---|---|---|
| Input | Data entry points (Text Prompt, Image Upload, Template Preset) | Green |
| Transform | Data manipulation nodes (Prompt Enhancer, Style Config, Image to Text, Translate Text) | Purple |
| Generate | AI-powered content creation (Image Generator, Video Generator, Background Remover, Inpainting) | Blue |
| Compose | Layout and composition tools (Text Overlay, Frame Border, Sticker Layer, Collage Layout) | Orange |
| Output | Final output nodes (Preview, Export) | Red |
- Node.js 18.x or higher
- pnpm 8.x or higher
- Backend server running (see
apps/server)
# Install dependencies (from monorepo root)
pnpm install# Start development server
pnpm devThe application will be available at http://localhost:4321.
# Create production build
pnpm build
# Preview production build locally
pnpm preview| Template | Description | Use Case |
|---|---|---|
| Ramadan Wishes | Generate personalized Ramadan greeting cards | Eid al-Fitr and Ramadan celebrations |
| Holiday Meme | Create viral holiday-themed memes | Social media content for holidays |
| AI Pet | Generate stylized pet character images | Personal pet avatars and stickers |
| Custom Avatar | Create personalized character avatars | Profile pictures and digital identities |
| Blank Canvas | Start from scratch with an empty workflow | Custom pipeline creation |
The application implements a client-orchestrated, server-executed architecture:
- Frontend Orchestrator: Manages execution graph topology, state management, and execution ordering
- Server Executor: Stateless AI processing endpoint that validates inputs and calls AI APIs
- Real-time Updates: SSE-based progress updates during pipeline execution
The port type system ensures data compatibility through a static compatibility matrix:
text: Plain text dataprompt: AI prompt data (enhanced text for generation)image: Image data with URL referencevideo: Video data with URL referencestyle: Style configuration (art style, color palette, mood)media: Generic media type (compatible with image and video)
The web application connects to the backend server. Ensure the following configuration:
API_BASE=http://localhost:3000
AI model configurations are defined in src/components/canvas/config/modelOptions.ts and can be customized based on available Alibaba Cloud Model Studio services.
ADIS AI leverages Alibaba Cloud's Qwen language model family to power intelligent content generation workflows. Qwen provides state-of-the-art natural language understanding and generation capabilities optimized for multilingual contexts, including Indonesian language support.
| Capability | Usage in ADIS AI |
|---|---|
| Text Generation | Create original content, prompts, and narratives |
| Prompt Enhancement | Intelligently enhance user prompts for better generation quality |
| Language Understanding | Analyze and process user inputs with cultural context awareness |
| Reasoning & Logic | Support complex workflow logic and conditional content generation |
| Multilingual Support | Native support for Indonesian and English language processing |
- Prompt Enhancement Node - Uses Qwen to analyze and enhance user prompts with creativity and tone modulation
- Template Presets - Qwen generates contextually relevant content for cultural templates (Ramadan, Holidays, etc.)
- Text Transformation - Text translation, rewriting, and style transformation capabilities
- Style Analysis - Intelligent analysis of style requirements and color palette suggestions
The application supports multiple Qwen model variants for different use cases:
- Qwen-Long - For processing extended context and complex reasoning
- Qwen-Turbo - Optimized for speed and real-time applications
- Qwen-Max - Maximum quality and instruction-following capabilities
Configuration of the active model can be adjusted in the application settings to match performance and accuracy requirements.
- Follow established code quality standards
- Maintain type safety and proper error handling
- Follow the existing component structure and naming conventions
- Add appropriate documentation for new features
- Test pipeline execution with various node configurations
This project is proprietary software. All rights reserved.
- Alibaba Cloud - AI infrastructure and model services
- Qwen - Advanced language model capabilities
- Open source community for visualization libraries