What We Shipped

A new Virtual Try-On Video tool creates a video of a model wearing your garment, so you can show how a piece moves rather than just how it sits in a flat photo.

• Multiple garments per video: dress a model in more than one item in a single clip
• Runs in the background: generation happens as a background job, so you can keep working while it renders
• Built for fashion catalogs: turn a static garment shot into a moving try-on without a real shoot

A new Marketplace Optimizer checks a product image against marketplace requirements and tells you whether it will pass before you list it.

• Compliance scorecard: see at a glance how an image measures up to the rules for marketplaces like Amazon and Shopify
• Suggested fixes: get concrete optimisations for what is holding the image back
• Catch problems early: review an image against the rules before upload instead of after a rejection

The new Bundle Composer generates product-bundle images, combining several products into a single image.

• Combine multiple products: bring 2 to 6 separate product images together into one bundle visual
• Two ways to run it: use it as a standalone tool or as a node inside a workflow
• Good for sets and kits: build the hero image for a multi-item listing without a manual mockup

You can now wire Sirv AI Studio into your automation tools and run AI operations without opening the app. Both connectors authenticate with your account API key.

• Zapier integration: trigger background removal and replacement, upscaling, image and video generation, alt text, object removal, depth maps, image-to-3D, and product lifestyle shots from any Zap
• n8n community node: call the same core operations (background remove and replace, upscale, image and video generation, alt text, object removal, depth maps, image-to-3D, and product lifestyle shots) from an n8n workflow
• Fits existing pipelines: drop a Sirv step into a larger automation so processed images flow straight into the rest of your stack

The MCP server now works inside ChatGPT through the OpenAI Apps SDK. You get an interactive widget for running AI tools without leaving the chat.

• Works in ChatGPT: an interactive widget runs background removal, upscale, background replace, and object removal directly in the ChatGPT UI
• No change for existing clients: Claude, Cursor, and other MCP clients keep working exactly as before
• One server, more places to use it: the same MCP server now reaches ChatGPT users alongside your editor and chat clients

Connect your Shopify store and process product images directly. The embedded app installs from the Shopify App Store and syncs your product catalog.

• Shopify embedded app with OAuth flow
• Sync product images for batch processing
• Process and push updated images back to Shopify

A new Image Translation tool translates the text inside an image into another language while keeping the original layout, so you can localise product graphics and labels without redoing the design.

• Layout stays intact: translated text replaces the original in place, preserving the fonts, colours, and positioning of the image
• Two ways to run it: use it as a standalone tool or drop it in as a step inside a workflow
• Built for product graphics: handles labels, badges, and on-image copy that would otherwise need a manual redraw per language

Sirv AI Studio is now available as an MCP server. Connect it to Claude, Cursor, or any MCP-compatible client and use all the AI tools directly from your editor or chat.

• Stateless HTTP mode for easy deployment
• OAuth token pass-through for authenticated access

Batch processing got a major upgrade under the hood and in the UI.

• Fan-out pattern for large jobs (100k+ items) using Inngest step parallelism
• Live preview of currently processing images via SSE stream
• First 10 images preview before the job starts so you can verify the selection
• Status-sectioned grid groups results by state (queued, processing, done, failed)

Import image URLs from CSV, Google Sheets, or XLS files to run batch upscale and alt-text jobs. Drop a spreadsheet, pick the URL column, and process everything in one go.