Status: [UNTESTED] Endpoints found: http://localhost:8081/api/swagger-ui.html http://localhost:8081/api/v1/tenants Source: https://github.com/suplab/aether-grid # Aether Grid > Distributed intelligence mesh — enterprise AI agent governance platform **Aether Grid** is the enterprise layer of the [Aether ecosystem](https://github.com/suplab/aether-core#ecosystem). It sits as a smart proxy and governance layer in front of any API ecosystem, orchestrating a mesh of specialised AI agents that learn, self-improve, and enforce policy across all API interactions. **Sister repository:** [suplab/aether-core](https://github.com/suplab/aether-core) — the personal cognitive engine that powers individual user context enrichment in Grid agents. **Author:** Suplab  |  **Status:** All 15 Phases Complete — Production Ready  |  **Phases Complete:** 0–15 of 15 --- ## What Aether Grid Does Aether Grid is implemented as a **Maven multi-module Spring Boot 3.x / Java 21** application. It operates as: - **Remembers** every API interaction semantically (all-MiniLM-L6-v2 embeddings + metadata in pgvector, 384-dim) - **Learns** patterns of successful and failing requests over time (PROCEDURAL/SEMANTIC/EPISODIC/EMOTIONAL memory types) - **Governs** API usage via YAML policies stored in PostgreSQL and evaluated at runtime with SpEL - **Predicts** temporal failure and latency windows with the `TemporalPredictionAgent` - **Debugs itself**: `HallucinationDetectorAgent` validates LLM outputs, `ReflectionAgent` evaluates system health - **Improves itself**: `SelfImprovingAgent` reads `AgentFeedback` history and generates LLM-powered improvement suggestions; `AgentLearningService` reviews all tenants weekly on a schedule - **Tracks performance**: `AgentController` exposes feedback recording (`POST /api/v1/tenants/{tenantId}/agents/feedback`) and per-agent performance stats (`GET /api/v1/tenants/{tenantId}/agents/performance`) - **Coordinates** six specialized agents across a cognitive mesh with shared pgvector memory - **Provides visibility**: live dark-theme operator dashboard at `http://localhost:8081/dashboard.html` —