Infinispan
infinispan.org
- Free plan
- Linux
- API
- Self-hosted
What Infinispan is
Infinispan is a free, Apache 2.0-licensed, open-source distributed in-memory database for applications that need caching alongside full-text and vector search. It supports distributed vector indexes, k-nearest-neighbor similarity queries, semantic search, retrieval-augmented generation, and semantic caching. Applications can connect through Hot Rod, REST, Redis, or Memcached; listed client libraries include Java, Go, JavaScript, C++, .NET, and Python. Run it as an embedded library, standalone server, or container image, with deployment options including a Kubernetes Operator and container images on OpenShift, AWS EC2, Azure, and GCP. Cache stores include filesystem, JDBC, SQL, RocksDB, and remote Infinispan clusters. The feature list includes access controls, security realms, encryption for inter-cluster traffic and remote endpoints, and distributed transactions with configurable isolation and XA recovery. Native integrations include Spring AI, LangChain4j, LangChain, and Quarkus LangChain4j. Its support policy covers only the latest minor release in a major version, with different levels of support.
Who it suits
Infinispan suits developers and teams building applications that need distributed caching, full-text or vector search, and flexible deployment. It may also fit teams using the listed client libraries, integrations, or cache stores.
Good to know, and what to watch
- Free under the Apache 2.0 license.
- Offers full-text search and distributed vector indexes.
- Supports four connection protocols.
- Runs embedded, standalone, or as a container image.
- Includes configurable distributed transactions and access controls.
- Only the latest minor release per major version is supported.
- Support levels vary by release and component.
Plans
Every plan and priceCommunity Edition
Free plan
The verdict
Infinispan combines distributed in-memory storage with search, multiple deployment modes, and a broad set of listed client options. Check its release support policy against your maintenance needs before choosing it.
What the maker publishes
- Client libraries
- The Hot Rod client page lists Java, Go, JavaScript, C++, .NET, and Python clients. source
- Cloud support
- The site lists a Kubernetes Operator and container image deployment on OpenShift, AWS EC2, Azure, and GCP. source
- Deployment
- Infinispan can run as an embedded library, standalone server, or container image. source
- Integrations
- The features page lists native integrations with Spring AI, LangChain4j, LangChain, and Quarkus LangChain4j. source
- License
- Infinispan is released under the Apache 2.0 open-source license. source
- Persistence
- Listed cache stores include filesystem, JDBC, SQL, RocksDB, and remote Infinispan cluster stores. source
- Product
- Infinispan is an open-source distributed in-memory database with full-text and vector search. source
- Protocols
- Clients can connect through Hot Rod, REST, Redis, or Memcached protocols. source
- Search and AI
- Features include distributed vector indexes, kNN similarity queries, semantic search, RAG, and semantic caching. source
- Security
- The feature page lists role and permission based access control, security realms, and encryption for inter-cluster traffic and remote endpoints. source
- Support channels
- For usage questions, the site recommends GitHub Discussions and Stack Overflow; development chat is on Zulip. source
- Support policy
- The releases page says only the latest minor release of a major version is supported and distinguishes full, limited, and limited client support. source
- Transactions
- Infinispan supports distributed transactions with optimistic or pessimistic locking, configurable isolation levels, and XA recovery. source
Checked against the maker's pages on 2 Oct 2026.