What We Shipped

A round of small fixes across the dashboard.

• Context menu first-click delay eliminated — menus now open instantly instead of requiring a second click after the first
• Vendor upload modal no longer causes infinite re-render when opened in certain states
• Empty library state now shows an "Upload files" button instead of a blank page
• Collection sidebar names no longer truncate — reduced padding so long names fit
• Notification dropdown layout fixed for long notification text
• Folder bulk selection in context menus now works correctly
• Generation tool links route to /create studio page instead of the old standalone page
• Workflows shortcut added to the studio mode dropdown for quick access

Cleaned up the Create Studio controls for a more focused workflow.

Mode tabs now use a compact dropdown with icons and labels instead of a horizontal tab bar that was getting crowded with five modes.

Create mode puts the prompt front and center. The generate button, model selector, and aspect ratio are all above the fold. Advanced settings (number of images, seed, negative prompt) are tucked below.

Lifestyle mode replaces the chip soup for scene selection with clean Select dropdowns for categories and placement, plus a tidy 2-column preset grid.

Video mode now supports multi-image reference for the veo31ref model. You can upload up to 5 reference images that get uploaded in parallel when you hit generate.

Completely reworked how you review, compare, and act on asset versions and AI derivatives inside the asset modal.

Version filmstrip sits below the main image viewer. Thumbnails for every version, keyboard arrow navigation, and right-click context menus for quick actions (compare, restore, download, delete). The current version is always pinned first.

Before/After comparison works for both versions and derivatives. Click any past version or hit "Compare" on a derivative, and a slider overlay appears in the main viewer area so you can scrub between original and processed.

Derivative actions are now front and center. Each derivative shows its operation name, dimensions, file size, and format. Hover to get action buttons:

• Compare opens the before/after slider against the original
• Download saves the derivative file
• Use as version promotes the derivative to a new version of the parent asset
• Assign to product shows an inline slot picker (Main Image, Gallery, Lifestyle, Detail, Swatch) when the parent is linked to a product
• Delete removes the derivative

All actions show spinners during processing and toast messages on success or failure.

Version timeline in the sidebar History tab now uses a clean dot-and-label layout. Compare and Restore buttons appear on hover. The old flat list and preview modal are gone.

The MCP server now covers the full breadth of the platform. Connect it to OpenClaw or any MCP-compatible assistant, and manage your entire product catalog from WhatsApp, Telegram, or Discord — without ever opening the dashboard.

Picture this: you're on the train, a supplier just sent new product shots. You forward them to your OpenClaw bot on Telegram: "generate lifestyle shots and a hero image for these, create a product video, write descriptions, alt text, and meta tags, then push everything to Shopify." Done before your stop.

Or just tell your OpenClaw assistant to check every morning for new products without images, generate the full asset suite, write all the copy, and publish. Your catalog stays fresh without anyone touching the dashboard.

• Asset tools list, search, update metadata, generate alt text, bulk operations, import from URL
• Product management create, update, delete, search products, link assets, bulk actions
• Shopify integration check connection status, push images to Shopify, sync product catalogs
• Account tools usage statistics with operation breakdown, list organizations
• Bearer token auth API routes now accept bearer tokens alongside session cookies, so the MCP server authenticates directly without browser sessions

Major decomposition pass across the orchestrator and canvas code. Every file now sits under 200 lines, making the workflow builder easier to work on and test.

• Canvas nodes split from a single 800+ line file into individual node components (source, step, output, edge nodes)
• Config layer extracted into 17 focused config files with a shared CanvasToggleGroup component
• Pipeline builder decomposed into dedicated hooks (useCanvasLayout, useCanvasEdges, useCanvasSteps) and smaller components
• Execution and upload hooks split into useExecutorState, useDerivedExecutionState, useUploadLogic, and useUploadUIState
• Workflow hooks broken into useWorkflowCrud and useWorkflowGraph
• Studio shared components extracted: SourceImagePicker, StudioResultDisplay, StudioModeTabs, useStudioPanel hook
• Stale closure fix in video preload callback
• Test coverage added for new shared studio components and hooks

The products page got a major upgrade — better filtering, infinite scroll, slot completion tracking, and inline product creation.

• Searchable filter dropdowns for brand, product type, and tags. Type to narrow down options instead of scrolling through long lists
• Faceted filtering with new server-side filters for brand, product type, tags, and image status. A dedicated /api/products/facets endpoint powers the dropdowns with distinct values from your catalog
• Slot completion tracking see filled/total slots per product at a glance. Filter by "slots complete" or "slots incomplete" to find products that still need images
• Infinite scroll replaces the old fixed-limit list. Products load in pages of 50 as you scroll, powered by useInfiniteQuery
• Inline product creation add products directly from the list page without navigating away
• Slot template grid on product detail the detail page now shows the slot template with filled/unfilled positions, reusing the vendor slot grid component
• Bulk edit product type added alongside existing bulk brand/category/tags editing
• More source filters WooCommerce and SAP sources now appear in the source dropdown
• Tag badges in list rows product tags display inline with overflow count

Three new modes in the Create Studio, bringing the total to five: Generate, Transform, Background, Lifestyle, and Video.

Background mode swaps or removes the background from a product shot. Pick a source image, describe the new background (or leave it blank for transparent), and the AI handles the cutout and compositing.

Lifestyle mode places your product into a styled scene. Upload a product photo, describe the setting ("marble countertop, soft morning light"), and get back a lifestyle shot ready for your store or social media.

Video mode turns a still image into a short product video. Pricing is shown upfront based on duration and resolution, pulled from a shared pricing module.

All five modes now share a common set of components: source image picker, result display, and mode tab switcher. Switching modes preserves your selected image. You can jump straight into a specific mode from the DAM context menu with pre-loaded images via /create?mode=background&imageId=... URLs.

• History grid updated to show all five operation types with distinct badges
• Creations collection auto-created on first generation save
• Video save support the save-generation endpoint now accepts mimeType for video files
• Seed and negative prompt controls added to Generate mode alongside num_images (1/2/4)

Uploading a file that already exists in a folder now shows a conflict dialog instead of silently creating a duplicate.

• Three resolution options per file: keep both (auto-rename with a suffix), replace existing, or skip
• Apply to all checkbox for batch resolution when multiple files conflict
• Replace preserves history the existing asset is updated in-place via a new /api/assets/[id]/replace endpoint, so linked products and metadata stay intact
• Filename tooltips long filenames in the gallery now middle-truncate with the extension visible, and show the full name on hover
• Hotkey suppression keyboard shortcuts no longer fire through open dialogs

New /create page inside the dashboard with a split workspace layout for AI image generation and transformation.

Generate mode turns text prompts into images. Pick a model (Nano Banana 2, Flux2, Seedream, GPT Image), set your aspect ratio and resolution, type a prompt, and hit Generate. Results auto-save to your library. You can refine the output with follow-up edits without starting over.

Transform mode takes an existing image (upload, pick from library, or paste a URL) and transforms it with AI. Style transfer, background changes, artistic effects. Same model picker, same auto-save flow.

Both modes share a history grid at the bottom showing your recent generations and transforms. The workspace uses a viewfinder-style preview that updates as you change aspect ratio and resolution settings.

AI tools now run directly from the asset modal. Open any image, switch to the AI tab, and run background removal, upscale, or any other tool with a before/after comparison right there.

• Context menu shortcuts open the modal on the AI tab and auto-start the selected tool
• Bulk AI processing from the batch action bar. Select multiple assets, pick a tool, process them all at once
• Auto-save derivatives processed results save as derivatives linked to the original asset
• State recovery if you refresh the page mid-processing, the tool state restores from sessionStorage
• PDF preview thumbnails for translated document derivatives