SCX.ai
SCX.ai has launched a fully managed vector storage product that turns documents into queryable embeddings on sovereign Australian infrastructure. According to the product page, developers can create a vector store, upload documents, and run context-aware retrieval-augmented generation through an API, with native connec The service uses the E5-Mistral-7B-Instruct model to convert each chunk into a 4,096-dimensional vector, which is then indexed in an isolated store with configurable distance metrics. A natural-language query is embedded and matched against the store via semantic similarity search, and the top retrieved chunks are inje
