Broadbrand
Platform · Owned by Broadbrand

A catalogue's worth of imagery. Without another studio day.

Persaic is the AI production layer for product imagery and product video. Catalogue cutouts and the full 5-photo product set per SKU. Lifestyle staging into real environments. Clothing and accessories composed onto AI-generated models. Missing variants generated from sibling references. Hero video from a single still. In production today at Liquid Lighting, where it generates catalogue imagery across thousands of SKUs.

AI for product photos + product videos
Persaic wordmark

Catalogue imagery, lifestyle scenes and on-model shots. Without another studio day.

Persaic is the AI production layer for product imagery and product video. Catalogue cutouts. Hero, three-quarter, low-angle and detail crops for every SKU. Lifestyle staging into real environments. On-model imagery for apparel, accessories and wearables. Missing-variant generation from sibling references. Hero video built from a single still. In production today at Liquid Lighting, where it generates the catalogue imagery library across thousands of SKUs.

Visit persaic.aiBook a demo
01
Product hero generation

Hero, three-quarter, low-angle and detail crops per SKU. Consistent framing, scale and lighting across the whole catalogue. Built for category breadth: furniture, lighting, decor, FMCG, electronics, beauty.

02
In-situ scene staging

Curated environments plus your own customer-supplied photos. Products placed at correct scale and orientation, with ambient lighting matched to each scene. Indoor, retail floor, outdoor or lifestyle backdrops.

03
Clothing & on-model imagery

Apparel, accessories and wearables composed onto AI-generated models. Pose, body type, ethnicity, age and styling controlled per shot. Editorial, lifestyle or in-store backdrops. Outfit composition across a full campaign.

04
Missing variants + product video

Colour, finish, fabric or body-type variants generated from sibling references. Geometry, lighting and shadows preserved. Hero videos and dynamic scenes built from a single still via Veo.

Pipelines on the production line
Catalogue extraction
  • +PDF or scanned catalogues in, transparent product PNGs out
  • +Tesseract OCR + Claude Vision pulls SKU, name and copy from each page
  • +Magenta-key cutout removes backgrounds and hollow interior pockets in one pass
5-photo product set
  • +Hero, three-quarter, low-angle, detail-base, detail-shade per SKU
  • +Auto-routed to the best-fit scene from a curated library
  • +Six parallel workers for catalogue-scale batch jobs
Scene + customer-photo staging
  • +Scene staging: product first, scene second, anchored at correct scale
  • +Customer-photo staging: environment first, product second — phone shots become campaign-ready
  • +Category-aware prompting so a sofa, a lamp or a kettle each get the right scene treatment
Clothing & modelling system
  • +Synthetic models composed with apparel, accessories or wearables
  • +Body type, ethnicity, age, pose and styling controlled per shot
  • +Multi-look outfit composition across a campaign, editorial or in-store backdrops
Variant generation
  • +Donor sibling image as anchor, target colour, finish or fabric as instruction
  • +Preserves exact geometry, lighting, shadows and detail
  • +Review queue with approve / reject before push to live

Works with: Gemini 3.1 Flash Image Preview, Veo, Claude Vision, Tesseract OCR, MinIO, Shopify, WooCommerce, Headless commerce, Akeneo PIM, Salsify, Cloudinary, Meta, Google Shopping.