Redis

Vector Search with Redis

Redis Vectors is an advanced feature within Redis, an open-source in-memory data structure store. Redis Vectors enables the use of vector similarity queries and supports real-time vector indexing, updates, and deletions.

Vector fields in Redis allow for the storage, indexing, and querying of vectors, either within Redis hashes or in JSON documents through the integration with the JSON module. This provides flexibility in organizing and accessing vector data.

The vector similarity functionality in Redis Vectors offers several key capabilities:

• Real-time Vector Indexing: Redis Vectors supports two indexing methods. The FLAT indexing method utilizes a brute-force approach, while the HNSW (Hierarchical Navigable Small World) indexing method is an optimized implementation of efficient and robust approximate nearest neighbor search.
• Real-time Vector Update/Delete: Updates or deletions of vectors trigger an automatic update of the corresponding index, ensuring data consistency.
• K-Nearest Neighbors (KNN) Search and Range Filter: Redis Vectors enables KNN search and range filtering based on distance metrics. Currently, three distance metrics are supported: L2 (Euclidean distance), IP (inner product), and HAM (Hamming distance).

Redis Vectors empowers developers and data professionals to efficiently work with vector data in Redis. By leveraging this feature, they can build applications that require similarity searches, recommendation systems, anomaly detection, and other vector-related use cases.

To learn more about Redis Vectors and how to utilize vector fields and vector similarity queries, please refer to the Redis Vectors documentation.

Aldo check Redis Modules:
https://redis.io/resources/modules/