Actors are lightweight and collaborate through asynchronous messaging for a common goal. Data can be shared by immutable messages.
Supports three different execution models — thread-bound actors, classic actor scheduling, or Loom-based virtual thread actors. GraalVM enables polyglot language support, here integrated through sandboxed native function pods.
Leverages the actor model to realize cognitive AI agents as encapsulated native function pods, providing a high-performance runtime for complex multi-agent orchestration.
The current version is 2.4, and it is still under further development. The Actor4j - Core library has no external dependencies and is also compilable as a native image with GraalVM. The minimum requirement is currently Java 21. The Actor4j - Core library is now separated into SDK and Runtime to provide more runtimes. Please note that the documentation is currently incomplete and does not yet fully reflect all the features implemented. New research publication available — Fall 2025!