A deterministic, Java 21 native AI runtime platform that brings intelligent capabilities directly into your JVM applications with predictable behavior and full observability.
Ten comprehensive client surfaces delivering production-hardened AI capabilities with fail-closed security and pgvector hybrid retrieval.
client.memory() — Persistent episodic context & recall
client.knowledge() — Hybrid RRF vector search & ingestion
client.planning() — Structured topological DAG planning
client.reasoning() — Deterministic thought & evidence graph
client.reflection() — Adaptive calibration & analytics
client.inference() — Structured scoring & tradeoff analysis
client.identity() — Tenant boundaries & profile resolution
client.execution() — Fail-closed action execution gate
client.project() — JavaParser Java 21 AST intelligence
client.settings() — Dynamic BYOK credentials & configuration
Add the Shree AI OS library to your Java 21+ project and start building deterministic intelligent applications immediately.
Add io.github.darshanrathod04:shree-ai-os:1.0.6-developer-preview to pom.xml
Bootstrap via ShreeAI.builder().apiKey(...) with BYOK support
Run grounded chat, episodic memory recall, and DAG planning
<!-- Maven Dependency -->
<dependency>
<groupId>io.github.darshanrathod04</groupId>
<artifactId>shree-ai-os</artifactId>
<version>1.0.6-developer-preview</version>
</dependency>
// Java 21 LTS Quickstart Bootstrap
import com.shreeai.os.platform.sdk.ShreeAI;
import com.shreeai.os.platform.sdk.SDKResponse;
import com.shreeai.os.platform.core.RuntimeConfiguration;
public class Quickstart {
public static void main(String[] args) {
// Configure runtime with fail-closed security & multi-kernel settings
RuntimeConfiguration config = RuntimeConfiguration.builder()
.defaultProvider("gemini-3.6-flash")
.failClosed(true)
.build();
// Initialize with API key or deterministic in-memory fallback
ShreeAI shree = ShreeAI.builder()
.apiKey(System.getenv().getOrDefault("GEMINI_API_KEY", "local"))
.configuration(config)
.build();
// Execute grounded chat through the 11-stage cognitive pipeline
SDKResponse response = shree.chat("Explain hybrid RRF vector retrieval");
System.out.println(response.answer());
shree.close();
}
}