Marqo is an AI-native search and discovery platform for ecommerce, built on vector search and multimodal embeddings. The open-source project on GitHub has more than 5,000 stars; the company has since deprecated the self-hosted open-source edition and moved development to its hosted platform at marqo.ai, which serves online brands in fashion, beauty, electronics and home goods.
Marqo turns product catalogs into searchable vector indexes where shoppers can search with text or images and get results ranked by semantic relevance rather than exact keyword matches. It layers personalization on top, using clickstream, purchase and event data to understand shopper intent, and automates much of the ranking and merchandising work that commerce teams traditionally do by hand. The goal is measurably commercial: better relevance, higher conversion and higher average order value. The original open-source release was notable for bundling its own vector search engine with inference for open embedding models, which is why the repository is still widely starred and forked even though the company now directs new users to the hosted platform.
Freemium in structure: the deprecated open-source edition remains available on GitHub at no cost, while the current, supported Marqo platform is a commercial hosted product sold to ecommerce brands. Check marqo.ai for current plan details.
Ecommerce and retail teams whose on-site search is a revenue lever, and who want AI-powered relevance and personalization without building a vector search stack themselves.