The latest features, improvements, and fixes to HuHu AI.
Users can now select a reference face directly from the Model Gallery, enabling more consistent identity across generated outputs.
Generation workflow to support creating targeted scenes and multiple poses with a selected model face in a single run.
The preview panel and generation history, making it easier to review outputs and track previous generations.
Jewelry as a supported input type in Bulk Try-On. Users can now upload jewelry SKU images and generate on-model visuals.
Supported jewelry types include necklaces, bracelets, watches, earrings, and rings.
Accessory support across Bulk Try-On, including items such as bags, glasses, hats, and scarves.
Multiple aspect ratio support for Bulk Try-On generation, including 1:1, 3:4, 4:5, and 9:16 formats.
4K output support for higher-resolution image generation.
AI-powered image quality assurance, which automatically reviews generated results and highlights the best image with a Recommended badge. Each recommendation includes a short explanation, and users can provide feedback on its usefulness.
Intercom live chat for real-time user guidance, bug reporting, and support.
API documentation, now available on the website.
Single Try-On mode, allowing users to generate virtual try-on images one by one by uploading a garment image and selecting a model, speeding up try-on image generation process when adding SKUs without managing files.
Bulk Video Generation mode, enabling users to upload multiple SKU images and generate video clips for each item in a single workflow.
Added video clip regeneration, allowing users to regenerate individual clips during the video generation process.
Model images can now be grouped by Model ID or Scene ID.
Added Model ID attributes to help manage model categories and sizes.
Automatic detection of view types (e.g., front/back) and crop types (e.g., full body, upper body, lower body) to support automatic model pairing in projects.
Model images can now be edited or used for pose changes directly from the Model Gallery.
Added the ability to update the cover image for a Model ID.
Multi Clip Mode lets you create up to 3 connected clips in one video generation — control camera movement and garment motion per clip for more dynamic results.
New Dynamic Intensity setting gives you control over how much movement appears in generated videos — from subtle fabric sway to full-motion scenes.
Try-On engine optimized to better preserve original garment stitching, texture, and layering details — especially for complex or multi-piece outfits.
New Overview Page lets you view all generated content — model images, try-on results, videos, and edits — in one unified feed.
Model images and videos now support Bulk Generation with the same workflow as try-on: add to queue, configure, and generate at scale.
Face Reference now works in both Model and Try-On generation — apply the same face across all content types for brand consistency.
Background Reference for Model Generation — upload a background image and the model will be placed naturally into that scene, matching perspective, lighting, and color tone.
Manual Model Pairing in Bulk Try-On — you can now assign specific models to specific garments before generating, giving full control over pairing.
Smart Generation Mode in Bulk Try-On — enable it to skip garments that have already been generated with the same model, saving credits and avoiding duplicates.
Model gallery now shows model usage status and supports filtering by paired/unpaired, making it faster to find and manage models across large projects.
All generation results now appear in a reviewable task list — approve, reject, or regenerate results before exporting.
Task panel redesigned to show status indicators, model-garment pairing info, and quick actions in one view.
Try-On quality improved for garments with fine details — lace, mesh, embroidery, and print patterns now render more accurately.
Video generation speed increased by ~30% with no reduction in output quality.
All model generation outputs are now 2K resolution by default — sharper detail for face, fabric, and accessories.
Model anatomy and hand quality significantly improved — fewer artifacts in fingers, wrists, and complex poses.
Lighting and shadow consistency enhanced across model poses to reduce unnatural hotspots and flat areas.
Garment Detail Control lets you fine-tune how closely the try-on result matches the original garment's texture, pattern, and structure — available for both single and bulk try-on.
Video Editing module — trim, crop, adjust playback speed, and add text overlays to generated videos directly in HuHu Studio.
Batch export for videos — download all generated videos for a project in one click with customizable naming and resolution settings.
Bulk Try-On now supports up to 500 garments per batch, with parallel processing and a real-time progress dashboard.
Garment pairing interface redesigned with drag-and-drop support — pair models with garments visually, with instant preview thumbnails and undo/redo.
Team workspace — invite team members to collaborate on projects with role-based permissions (Admin, Editor, Viewer).
Activity log tracks all generation, editing, and export actions per team member.
Project organization upgraded with folders, tags, and search — find any asset across all projects instantly.
Dashboard now shows team usage stats, credit consumption, and generation history in a single view.
Export options expanded — download in PNG, JPEG, WebP, or TIFF with configurable quality and metadata settings.
Auto Batch Mode — upload a product catalog CSV and HuHu automatically creates try-on tasks for each SKU, matching garment images to models based on your rules.
20+ new preset models added to the model gallery — diverse body types, ethnicities, and poses optimized for fashion e-commerce.
Multiple Garment Input — combine up to 4 garments (top, bottom, outerwear, accessories) in a single try-on to create complete outfit looks on one model.
HuHu AI Team Version launched — centralized workspace for brands and studios with shared model libraries, project templates, and team billing.