DiskANN is now generally available in vCore-based Azure Cosmos DB for MongoDB, giving you powerful vector search capabilities directly within your database. This means you can now build intelligent applications that search across high-dimensional vector data—like text embeddings, images, or other AI-generated features—with low latency and high accuracy.
With support for up to 4,000 dimensions, this release makes it easier to work with modern embedding models and large-scale AI workloads. DiskANN uses an optimized compression algorithm (PQ) to store and search vectors efficiently, reducing memory consumption while boosting search speed.