Group: Akilisha OSS

Sort by:Popular▼

✓Popular✓Newest

The Express-inspired foundation of CafeAI — app bootstrap, routing, the middleware chain, and the AI primitives over Helidon SE.

Safety and regulatory guardrails for CafeAI — PII scrubbing, jailbreak and prompt-injection detection, toxicity filtering, topic boundaries, and regulatory compliance.

HTTP identity for LangChain4j AiServices agents — session threading, guardrail pre-screening, and observability, bound at AiServices.builder() time.

HTTP identity for LangChain4j AiServices agents — session threading, guardrail pre-screening, and observability, bound at AiServices.builder() time.

Tiered context-memory strategies for CafeAI — in-JVM, off-heap SSD, and Redis.

Durable workflow graphs for CafeAI -- steps, branches, loops, parallel branches, signals and timers, kept in a StateStore.

Bill of materials for CafeAI: aligns every published CafeAI module on one version.

HTTP identity for langchain4j-agentic multi-agent workflows -- the app's model, guardrails, and observability pre-wired into every agent.

The cafeai command: sign in to a model once.

File-based application configuration for CafeAI — application.properties/.yaml with environment profile overlays, layered under cafeai-core's AppConfig.

Out-of-process service connectivity for CafeAI — Redis, Ollama, and pgvector.

Background agent tasks for CafeAI -- child agents, approvals, budgets, checkpoints and sandboxed commands around LangChain4j AiServices.

Development-time reload for CafeAI apps: save, and the new code is live.

CafeAi examples module

JVM-level visibility for CafeAI via Java Flight Recorder, surfaced as OpenTelemetry metrics.

Verified caller identity for CafeAI from any OpenID Connect issuer: bearer tokens, sign-in flows, and credentials for model calls without API keys.

Serves CafeAI routes and @Tool objects as MCP tools for AI agents.

Optional Mustache template engine for CafeAI views.

Tracing for CafeAI, built on OpenTelemetry, with a console strategy for local development.

Retrieval-Augmented Generation for CafeAI — document sources, embedding providers, and the PgVector and Chroma vector stores.