> ## Documentation Index > Fetch the complete documentation index at: https://docs.wrldwide.ai/llms.txt > Use this file to discover all available pages before exploring further. # 节点系统 > 可视化工作流构建器 — 连接 36 种 AI 节点类型，构建自动化创意流水线 # Node System 使用拖拽式可视化工作流构建器构建复杂的 AI 创意流水线。跨输入、处理、逻辑和输出类别连接 **36 种节点类型**，构建自动化工作流。支持即时运行、定时调度或批量处理模式。 ## 概览 Access the Node System at `/nodes` to create and manage workspaces. Each workspace is a canvas where you connect nodes to build workflows. ### Key Benefits * **Visual Programming** — No code required, drag-and-drop interface * **Chain Operations** — Connect multiple AI models in sequence * **Conditional Logic** — Use If/Else nodes for dynamic workflows * **Batch Processing** — Process hundreds of items with List nodes * **Reusable Templates** — Save and share workflows * **Real-Time Preview** — See outputs as nodes execute ## Node Categories ### Input Nodes **Prompt Node** * Enter text prompts for AI generation * Supports variables and dynamic text * Can be connected to multiple processing nodes **Image Upload Node** * Upload images from device * Supports drag-and-drop * Outputs image URL for downstream nodes **Video Upload Node** * Upload videos for processing * Used with video-to-video models * Motion transfer, style transfer **Audio Upload Node** * Upload audio files * Used with lip sync, AI avatar * Supports MP3, WAV formats **Number Node** * Input numeric values * Used for parameters (strength, duration, etc.) * Can be used as variables **Style Node** * Predefined art styles * Applies style keywords to prompts * Photographic, Artistic, 3D, etc. ### Processing Nodes **Model Node** (Image Generation) * Generate images with any of 19 image models * Select model, resolution, aspect ratio * Connect to Prompt/Style nodes **Video Generate Node** * Create videos with 28 models * Text-to-Video or Image-to-Video * Duration (5-15s), aspect ratio, audio options **Cinema Studio Node** ✨ NEW * Professional cinema controls * 12 cinema lenses (spherical/anamorphic) * Focal length, aperture, 6-axis camera movement **Motion Control Node** ✨ NEW * Transfer motion from reference video * Apply motion patterns to any image * Precise motion replication with Kling 2.6 **Upscale Node** * Enhance image quality * Recraft Crisp (4x) or Topaz Pro (4x) * Supports up to 16MP input **Edit Image Node** (Inpainting) * Modify images with Flux Kontext * Pro (2 credits) or Max (4 credits) modes * Paint mask to edit specific areas **Remove Background Node** * Professional background removal * Clean cutouts for products * Transparent PNG output **Relight Node** * AI-powered relighting with 18 styles and 7 light directions * 3D depth-aware cinematic relighting * Control direction, color, brightness, and quality **Face Swap Node** * Replace faces in images/videos * Upload source face * Seamless blending **Lip Sync Node** * Sync lips to audio * Upload video + audio * Natural mouth movements **Live Avatar Node** * Create talking avatars * Photo + audio → animated video * Real-time lip sync **Video Compositor Node** * Combine and composite multiple videos * Layer video elements together * Professional video compositing **Speed Ramp Node** * Apply speed ramp effects (slow-mo, fast-forward, ramp profiles) * Dynamic speed control for video * Cinematic speed transitions **Image 3D Node** * Convert images to 3D models * GLB, USDZ, FBX export formats * WebGL preview **Assistant Node** * AI chat integration * Prompt enhancement * Text transformation **Translate Node** * Translate text between languages * 100+ languages supported * Preserves formatting ### Logic & Utility Nodes **If/Else Node** * Conditional branching * Route data based on conditions * Example: Filter by quality score **Filter Node** * Filter results by criteria * Remove unwanted outputs * Score-based filtering **Merge Node** * Combine multiple outputs * Merge images, videos, or data * Used for comparisons **List Node** * Batch process multiple inputs * Loop through prompts, images * Parallel execution **Aspect Ratio Node** * Resize/crop images * Change aspect ratio * Maintain or crop to fit **Social Share Node** * Post directly to social media * Connect social accounts * Auto-publish workflows ## Building Your First Workflow ### Simple Image Generation ``` [Prompt Node] → [Model Node] → [Output] ``` 1. Add **Prompt Node**, enter "A futuristic city" 2. Add **Model Node**, select FLUX 2 Pro 3. Connect Prompt → Model 4. Click **Run Workflow** ### Image Generation + Upscaling ``` [Prompt] → [Model Node] → [Upscale Node] → [Output] ``` Generate at 1K, upscale to 4K for maximum quality at lower cost. ### Multi-Model Comparison ``` [Prompt] → [Model Node (Flux)] → [Filter] → [Model Node (Grok)] → [Filter] → [Merge] → [Output] ``` Test the same prompt across multiple models, merge results for side-by-side comparison. ### Batch Product Photos ``` [List Node: 10 products] → [Model Node] → [Remove Background] → [Upscale] → [Bulk Export] ``` Process entire product catalog with one workflow. ### Video + Face Swap Pipeline ``` [Prompt] → [Video Generate] → [Face Swap] → [Output] ``` Generate video, then swap in custom face. ## Advanced Workflows ### Conditional Quality Control ``` [Prompt] → [Model Node] → [If/Else: Score > 0.8?] ├─ YES → [Upscale] → [Export] └─ NO → [Retry with different model] ``` Automatically upscale only high-quality outputs. ### Multilingual Content Generation ``` [English Prompt] → [Translate to Spanish] → [Model Node] → [Output ES] → [Translate to French] → [Model Node] → [Output FR] ``` Generate localized content for multiple markets. ### A/B Testing Creative ``` [Base Prompt] → [Style: Photographic] → [Model] → [Filter] → [Merge] → [Style: Artistic] → [Model] → [Filter] → [Merge] ``` Test different styles, compare results. ### Product Animation Pipeline ``` [Product Image] → [Remove Background] → [Video Generate] → [Upscale Video] → [Export] ``` Clean product → animated video → enhanced quality. ## Node Features ### Variables Use `{variable_name}` in prompts to create dynamic workflows: ``` Prompt: "A {color} {product} in {environment}" ``` Connect Number/Text nodes to inject variables. ### Batch Processing List Node can process: * Multiple prompts (text array) * Multiple images (URL array) * Multiple parameters (number array) **Parallel Execution:** All items process simultaneously for speed. ### Error Handling Nodes handle errors gracefully: * **Retry Logic** — Auto-retry failed generations * **Fallback** — Use If/Else for backup models * **Skip** — Continue workflow even if node fails ### Caching Nodes cache results to save credits: * Rerunning same input = instant result, no charge * Clear cache to force re-generation ## Workflow Templates Pre-built workflows available in the template library: * **E-commerce Product Shots** — Remove BG + upscale + variations * **Social Media Content** — Generate + resize for all platforms * **Video Ads** — Image → Video → Face Swap → Export * **Logo Variations** — One prompt → 10 style variations * **Multilingual Posts** — Translate + generate localized images Access templates: Click **Templates** in Node System sidebar. ## Keyboard Shortcuts * **N** — Open node selector * **Delete** — Delete selected node * **Cmd/Ctrl + C** — Copy node * **Cmd/Ctrl + V** — Paste node * **Cmd/Ctrl + D** — Duplicate node * **Cmd/Ctrl + Z** — Undo * **Cmd/Ctrl + Shift + Z** — Redo * **Cmd/Ctrl + S** — Save workflow * **Space + Drag** — Pan canvas * **Mouse Wheel** — Zoom in/out ## Sharing & Collaboration ### Save Workflows 1. Click **Save Workflow** 2. Name your workflow 3. Add description and tags ### Share Workflows 1. Click **Share** 2. Generate shareable link 3. Optional: Publish to Marketplace ### Marketplace * Sell workflows to other users * Set credit price per workflow * Earn from popular templates See [Marketplace](/en/features/marketplace) for details. ## 使用场景 ### E-commerce * Bulk product photography * Seasonal campaign variations * Localized product shots ### Marketing * A/B test creative concepts * Multi-platform content (resize for all socials) * Campaign asset generation ### Content Creation * Batch YouTube thumbnails * Consistent brand assets * Series/episode artwork ### Development * API-free workflow automation * Prototype AI features * Test model capabilities ## 最佳实践 ### Optimize for Speed * Start with low resolution, upscale final outputs * Use batch processing (List Node) for parallel execution * Cache results to avoid re-generation ### Optimize for Cost * Generate at 1K, upscale to 4K (cheaper than direct 4K) * Use If/Else to only upscale best outputs * Reuse cached results when iterating ### Organize Workflows * Name nodes clearly ("Product Image Upload", not "Image Upload 1") * Group related nodes visually * Add comments with sticky notes ### Debug Workflows * Enable **Debug Mode** to see node outputs * Use **Workflow Debug Panel** to inspect data flow * Test nodes individually before chaining ## Limitations * **Max Nodes:** 100 nodes per workflow (contact support for limits increase) * **Execution Time:** 30 minutes max per workflow run * **Concurrent Runs:** 5 workflows can run simultaneously ## API Integration Execute workflows via API: ```bash theme={"dark"} curl -X POST https://wrldwide.ai/api/workflow/run \ -H "Authorization: Bearer YOUR_API_KEY" \ -d '{ "workflow_id": "YOUR_WORKFLOW_ID", "inputs": { "prompt": "A futuristic city" } }' ``` ## 下一步 Learn advanced batch techniques Sell your workflows Test models side-by-side Team collaboration features