Fabric Databases are based on Azure SQL Databases.

The support for vector search is included in both and it's a powerful feature to build AI architectures, either RAG or similar architectures depending on vector search features

Using HTTP calls from T-SQL and vector search features in SQL we can store embeddings inside Fabric Database. This provides a considerable improvement in text search and a potential integration between database solutions and Open AI solutions.

The embeddings stored in Fabric databases can be used as input for Open AI models with the benefit of the flexibel search of a relational database