One unified REST API and a 24-tool MCP gateway power AI-native search, data quality, agentic feeds, and virtual try-on; free for the first 500 developers.
visualAI retail solutions, inc today announced the launch of discoverGPT, an AI discovery platform built for the agentic commerce era. discoverGPT gives developers and agencies one credential and one canonical product catalog behind a unified REST API and a 24-tool Model Context Protocol (MCP) gateway, covering the full discovery loop: onsite AI-native search, product data quality, universal offsite agentic feeds, and virtual try-on.
Keyword search has not kept pace with how people shop. Shoppers describe what they want in natural language, search with images, and increasingly ask AI agents such as ChatGPT, Gemini, and Perplexity to find products for them. ChatGPT alone handles an estimated 50 million shopping queries a day, according to OpenAI research, and AI-referred traffic converts at roughly five times the rate of organic search, 14.2% versus 2.8%, per the Apollo AI Search Benchmark. Adobe’s Q3 2026 data shows AI-driven retail traffic up 138% year over year.
Marketing Technology News:Â MarTech Interview with Miguel Lopes, CPO @ TrafficGuard
“Discovery didn’t migrate, it doubled,” said Joe Monastiero, Founder and CEO of visualAI. “Shoppers still start on the storefronts they trust, and a merchant’s own site remains their highest-intent, highest-margin surface. But they’re also asking AI agents to find products, and most catalogs are invisible to those agents. discoverGPT gives developers one platform to win on both fronts, on the merchant’s terms, without handing their brand or their customer relationship to somebody else’s marketplace.”
discoverGPT connects a merchant’s catalog once and makes every capability available immediately, over the same API, against the same data:
– Search: trimodal product discovery combining natural language, precise color, and image similarity in one flow.
– Data Quality: AI-driven product data cleaning, enrichment, and completeness scoring that repairs bad titles and missing attributes at the source.
– Agentic Feeds and GEO: machine-readable feeds (JSON-LD, llms.txt, UCP/ACP) that make a catalog readable to agents like ChatGPT, Gemini, and Perplexity, with visibility tracking into how AI agents see the store.
– Virtual Try-On: AI image composition with digital twin technology that lets shoppers see products on themselves in seconds.
Marketing Technology News:Â Is the Traditional CDP Already Out of Date?
The platform is MCP-native from day one. Any MCP-aware host, including Claude, ChatGPT, Cursor, or a custom agent, can point at the discoverGPT gateway and discover all 24 tools automatically, with no custom integration code.
discoverGPT is platform-neutral by design. Catalogs can come from Shopify, Salesforce, Magento, BigCommerce, WooCommerce, custom PIMs, or flat CSV exports, all mapped into one canonical schema through a free in-browser Column Mapper.
The platform runs on more than 350,000 lines of AI-native infrastructure code, not a wrapper around a third-party model. The underlying engine is already live in production on real shopper traffic: shopperGPT, the trimodal search capability discoverGPT exposes over REST and MCP, is in the Shopify App Store and running on visualAI’s Founding Merchant Partner, Cassie’s Kurations, at cassieskurations.com, alongside the studioGPT virtual try-on extension. The core technology is patent pending (34 claims, USPTO Application 19/211,109).
The Starter tier is free for life for the first 500 developers, with full REST and MCP access, no credit card required. An API playground with a shared demo catalog lets developers make their first call in minutes.











