[可一键复制] ComfyUI 通用媒体加载器 · 图像提示词模板

提示词可在此修改后复制或实测，不会改动源内容

Hello, I got tired of cluttering my workflows with multiple different loader nodes for different inputs (Load Image, Load Video, VHS, Audio Loaders, Empty Latent calculators, etc.). So I built **ComfyUI_UniversalMediaLoader**, a single, unified node with a rich interactive UI that handles everything:

✨ Key Features:
- 📁 **Drop Anything**: Images (PNG/JPG/WEBP), Videos (MP4/MOV/WEBM), Animated GIFs, and Audio files (WAV/MP3/FLAC).
- 🎨 **Canvas Planner Mode**: When no file is loaded, use it as a visual resolution/aspect ratio planner (1:1, 16:9, 4:3, 3:2 + Landscape/Portrait toggle) with automatic 32px grid snapping and Megapixel clamping for SDXL/Flux latents.
- ✂️ **Visual Crop & Outpaint**: Interactive bounding box with aspect ratio locking. Pull it outside the image bounds to instantly generate outpaint masks & padding.
- 🖌️ **Inpaint Mask Brush**: Draw inpainting masks directly inside the node with brush/eraser and mouse wheel size control.
- 🎬 **Video & GIF Timeline**: Full timeline with trim handles, playhead scrubbing, custom FPS resampling, audio mute/loop, and frame extraction.
- 🎵 **Audio Waveform & Speed Scaling**: Waveform display, 1-second magnet trim snapping, and time-stretching with pitch preservation (0.10x to 4.00x).

📦 Modular Unpack Nodes:
- 📐 **Universal Size Unpack**: Feeds exact width/height/aspect ratios to Empty Latent nodes instantly without decoding heavy image/video tensors.
- 🖼️ **Universal Image Unpack**: Outputs RGB, RGBA, and inpaint masks. Connected to a Video or GIF, it automatically extracts a 3-frame batch [Start, Playhead, End].
- 🎧 **Universal Audio Unpack**: Outputs clean audio waveform dictionaries, exact sample rate, trimmed durations, channel count, and handles time-stretching.
- 🎞️ **Universal Video Unpack**: Decodes videos & GIFs into frame batches [B, H, W, C] with synchronized masks and audio tracks.

填写任务材料

填写材料后生成预览，材料仅在当前页面内处理。匹配的占位符会被替换，其余材料追加在末尾；请检查与原文是否一致。

media_file features