{"status":"AgentHub Backend Running","version":"2.3.0","features":["Multi-agent orchestration","Two-phase manager logic (analyze → plan → execute)","Trajectory replay","A/B testing with BERT + ROUGE + LLM eval","Human-readable experiment reports","🧠 Agentic Memory (PostgreSQL + pgvector)","💰 Agent Economics & Cost Governance","Full transparency (events)"],"memory":{"backend":"PostgreSQL + pgvector","features":["Episodic memory (past runs)","Semantic memory (facts & entities)","Semantic search (OpenAI embeddings)","Importance-based retrieval","Memory decay & forgetting","User preferences","Agent learning","Error pattern avoidance","Working memory (per-run context)"]},"cost_governance":{"features":["Real-time token & cost tracking on every LLM call","Per-task cost attribution (itemized receipts)","Budget enforcement (soft/hard limits)","Cost-aware orchestration (pre-execution estimates)","Agent cost leaderboard (which agents burn money)","Optimization recommendations (model downgrade, prompt compression)","What-if analysis (cost comparison across models)","Anomaly detection (3σ cost spike alerts)","Prompt efficiency scoring","ROI scoring (value per dollar, A+ to C- grades)","Cost Memory Loop (learns from every run)","Cost-aware planning (LLM reasons about cost)","Experiment-backed recommendations (A/B proven)","Cost trend alerts (proactive spending detection)","Cost Autopilot (one-click model optimization)"],"endpoints":["GET  /costs/pricing       - Model pricing reference","GET  /costs/estimate       - Quick cost calculator","GET  /costs/summary        - User cost summary","GET  /costs/runs/{run_id}  - Itemized run receipt","GET  /costs/agents         - Agent cost leaderboard","GET  /costs/optimize       - Optimization recommendations","GET  /costs/whatif/{run_id} - What-if model analysis","POST /costs/budgets        - Set spending budgets","GET  /costs/anomalies      - Detect cost spikes","GET  /costs/roi/{run_id}           - ROI grade","GET  /costs/trends                 - Spending alerts","GET  /costs/smart-recommendations  - Smart recs","POST /costs/autopilot              - Auto-optimize"]}}