What We Shipped

New AI tool that writes product descriptions from product images. Upload a photo and get a ready-to-use product listing description with key features, materials, and selling points.

Also in this release: Alt text batch mode — generate alt text for multiple images at once, and video model previews showing example outputs before you pick a generation model.

Stripe subscriptions are now production-ready with proper upgrade/downgrade flows, and a full security audit closed 20+ vulnerabilities.

• Pro-rated upgrades and downgrades with clear messaging about billing changes
• Dynamic pricing for upscale and video generation based on resolution and duration
• Credit rollover and pending transaction cleanup on plan changes
• Payment intent verification to prevent amount tampering
• Rate limiting with per-operation buckets
• Workflow ownership validation — users can only access their own workflow runs
• Path traversal prevention on Sirv uploads with filename sanitization
• SSRF protection — removed runtime mutation of trusted domain allowlist

Large batch jobs now run on the server via Inngest instead of in your browser. Close the tab and come back later — your pipeline keeps running.

• Auto-switch to server-side execution for batches over 10 images
• SSE progress updates — real-time status pushed to the browser via server-sent events
• Reconnect on refresh — if you leave and come back, the UI reconnects to the running job and picks up where it left off
• Cancel support — stop a running server job from the UI

New activity feed showing your recent AI operations, plus a fresh look for the tool pages.

• Activity page with a timeline of all your image generations, removals, upscales, and other AI operations
• Save to Library — send any processed result from the activity feed to your asset library with one click
• Image-to-Image redesign — cleaner layout with model explanations, drag & drop source upload, and better result display
• 3D model preview — inline GLB viewer on the Image-to-3D tool page
• Sirv upload status — real-time indicator showing when results are being synced to your Sirv CDN account

Sirv AI Studio now works as a Progressive Web App on mobile. Install it from your browser and use it like a native app.

• Pull-to-refresh on mobile with safe area handling
• Sticky action bars on tool pages for easy access on small screens
• Collapsible quick actions on the dashboard
• Fullscreen mode for the workflow orchestrator canvas
• Undo/Redo in the orchestrator with Ctrl+Z / Ctrl+Shift+Z and a 50-entry history stack
• Step duplication — clone any pipeline step with its full config

Chain multiple AI tools into automated pipelines. Drop nodes on a visual canvas, connect them, and run entire workflows on single images or batches.

• Visual canvas editor with drag-and-drop nodes
• Review nodes to pause the pipeline for human approval
• Simple mode for quick linear chains without the full canvas
• All existing tools available as pipeline steps

Upload a video and get auto-generated captions. Useful for social media clips and product demos where you need subtitles fast.

Pick files directly from your Sirv CDN account without downloading and re-uploading. Plus, sign in with your existing Sirv account.

• Sirv file picker to browse and select assets from your library
• Upload back to Sirv after processing, with custom path and options
• Sign in with Sirv as an auth option alongside GitHub and email

Buy credits with Stripe. Monthly subscriptions and one-time credit packs are both available. Email receipts send automatically, and your credit balance shows on every tool page.

• Follow-up edits on generated images
• Delete account option in settings
• Mobile horizontal scroll fix for iOS

Process multiple images at once. Select files, pick a tool, and let it run through the queue. Results show in a before/after compare slider so you can check quality at a glance.

• Batch mode for background removal, upscaling, and more
• Compare slider with side-by-side before/after view
• Multi-select in the dashboard gallery
• Progress tracking per image