Redis 8 Integrates Modules into Core

Redis turned a bundle of add ons into one default install, which makes Redis much easier to adopt as an all purpose real time data layer. Before Redis 8, teams often reached for Redis Stack or extra modules to get JSON documents, full text search, time series, and probabilistic filters. With Redis 8, those features ship in core, alongside the query engine and newer vector capabilities, so one server can handle caching, retrieval, filtering, and lightweight search in a single operational footprint.

• The practical change is in setup and operations. A developer no longer has to install and version RediSearch, RedisJSON, RedisTimeSeries, or RedisBloom separately. Redis 8 versions these capabilities in lockstep with the main database, which cuts integration work and reduces upgrade mismatch risk.
  1 redis 3 redis 4 redis
• This also sharpens Redis's pitch against both point tools and broader databases. A team building semantic search or RAG can store embeddings, attach JSON metadata, run hybrid text and vector queries, and use Bloom style filters in the same system instead of stitching together a cache, a document store, and a vector database.
  2 redis 5 sacra 6 sacra
• The integration is also part of Redis's licensing and packaging reset after the Redis Stack era. In May 2025, Redis paired Redis 8 with AGPLv3 and folded Stack technologies into core, which helped reposition the product as a single distribution rather than a base database plus an add on bundle.
  2 redis 7 redis

Going forward, Redis is moving further from being just a cache and closer to being the default fast path database for application state and AI retrieval. As more developers expect vector search, JSON, and filtering out of the box, the winning products will be the ones that remove assembly work and let one deployment cover more of the live application stack.