Marqo

Marqo is a vector search engine that bundles embedding generation and similarity search into a single service. Unlike vector databases that only store and query embeddings (and force you to bolt on a separate embedding API), Marqo accepts raw text, images, or video and handles tokenization, embedding, indexing, and retrieval in one call. That removes a large class of integration bugs and lets developers iterate on retrieval quality by swapping models inside Marqo rather than rewriting their data pipeline. The product comes in two flavors: an open-source self-hosted engine (Marqo OSS, available on GitHub and Docker) and Marqo Cloud, the fully managed offering. Marqo is well known for strong multimodal retrieval — search a fashion catalog by an image, or find video clips by a text query — and ships pre-trained embedding models for general, e-commerce, and fashion use cases, with the ability to fine-tune your own on top. Customers include e-commerce, content discovery, and enterprise search teams. Pricing for Marqo Cloud is contact-driven on the marketing site (the pricing URL was a 'Book a Demo' page during automated fetch), so specific plan rates need manual verification.