Update README.md

README.md

@@ -1,5 +1,5 @@
 ---
-title: Agentic Reliability Framework (ARF) v4 API
+title: Agentic Reliability Framework (ARF) v4 – Public API Demo
 emoji: 🤖
 colorFrom: blue
 colorTo: green
@@ -9,23 +9,25 @@ app_file: app.py
 pinned: false
 ---
 
-# Agentic Reliability Framework (ARF) v4.2.0 – AI Reliability & Self-Healing Control Plane
+# Agentic Reliability Framework (ARF) – Public API Demo (Sandbox)
 
 **Problem:** Most AI‑driven governance systems fail silently in production, leading to outages, security breaches, and compliance violations.
 
-**Solution:** ARF turns probabilistic AI into deterministic, auditable action using Bayesian inference, semantic memory, and deterministic probability thresholds (DPT).
+**Solution:** ARF turns probabilistic AI into deterministic, auditable action using Bayesian inference, semantic memory, and **expected loss minimisation**.
 
 **Outcome:** Reduce MTTR by up to 85% with self‑healing systems, backed by fully explainable risk scores.
 
+> ℹ️ **This Space provides a sanitised, mock API endpoint.** The real ARF core engine is proprietary, access‑controlled, and available only to qualified pilots and enterprise customers. See the [public specification](https://arf-foundation.github.io/arf-spec/) for details.
+
 ---
 
 ## 🚀 Start Here
 
 | | |
 |--|--|
-| **📚 API Docs** | [https://a-r-f-agentic-reliability-framework-api.hf.space/docs](https://a-r-f-agentic-reliability-framework-api.hf.space/docs) |
-| **🧪 Live Demo** | [Gradio Dashboard](https://a-r-f-agentic-reliability-framework-api.hf.space/) |
-| **📦 GitHub** | [github.com/arf-foundation/agentic-reliability-framework](https://github.com/arf-foundation/agentic-reliability-framework) |
+| **📚 API Docs** | [https://a-r-f-arf-sandbox-api.hf.space/docs](https://a-r-f-arf-sandbox-api.hf.space/docs) |
+| **🧪 Live Demo** | [Gradio Dashboard](https://a-r-f-arf-sandbox-api.hf.space/) |
+| **📦 Public Spec** | [github.com/arf-foundation/arf-spec](https://github.com/arf-foundation/arf-spec) |
 | **📅 Book a Call** | [Calendly](https://calendly.com/petter2025us/30min) |
 
 ---
@@ -36,20 +38,20 @@ pinned: false
 import requests
 
 response = requests.post(
-    "https://a-r-f-agentic-reliability-framework-api.hf.space/v1/incidents/evaluate",
+    "https://a-r-f-arf-sandbox-api.hf.space/v1/evaluate",
     json={
         "service_name": "payment-gateway",
         "event_type": "latency_spike",
         "severity": "high",
-        "metrics": {"latency_p99": 350, "error_rate": 0.12}
+        "metrics": {"latency_p99": 450, "error_rate": 0.12}
     }
 )
 print(response.json())
 ```
 
-The response includes a full HealingIntent with:
+The response includes a mock HealingIntent with:
 
-*   risk\_score: calibrated failure probability
+*   risk\_score: simulated failure probability
     
 *   risk\_factors: additive contributions from conjugate prior, hyperprior, and HMC
     
@@ -58,14 +60,16 @@ The response includes a full HealingIntent with:
 *   decision\_trace: expected losses and variance
     
 
-🧠 Key Capabilities
--------------------
+⚠️ **All responses from this endpoint are simulated.** The real Bayesian engine is not exposed publicly.
+
+🧠 Key Capabilities (Conceptual Overview)
+-----------------------------------------
 
 *   **Bayesian Risk Scoring** – Conjugate priors + HMC for calibrated uncertainty.
     
 *   **Semantic Memory** – FAISS‑based retrieval of similar past incidents.
     
-*   **DPT Thresholds** – Deterministic approve/deny/escalate (0.2/0.8).
+*   **Expected Loss Minimisation** – Chooses approve/deny/escalate by minimising cost-weighted risk, not static thresholds.
     
 *   **Multi‑Agent Orchestration** – Anomaly detection, root cause, forecasting.
     
@@ -76,7 +80,7 @@ The response includes a full HealingIntent with:
 ```text
 User Request → Policy Evaluation → Cost Estimation → Risk Scoring
                                                     ↓
-                           HealingIntent ← Decision (Bayesian Loss)
+                           HealingIntent ← Decision (Expected Loss)
 ```
 
 All decisions are immutable, signed, and fully traceable via ancestor\_chain and infrastructure\_intent fields.
@@ -94,4 +98,6 @@ Then open [http://localhost:7860](http://localhost:7860/) for the Gradio UI an
 📚 About ARF
 ------------
 
-Agentic Reliability Framework is an open‑source advisory engine for cloud infrastructure governance, powered by Bayesian inference. Learn more at [github.com/arf-foundation](https://github.com/arf-foundation).
+The **Agentic Reliability Framework** is a governed, mathematically grounded advisory layer for AI infrastructure. The public specification, demo UI, and sandbox API are open‑source (Apache 2.0). **The core Bayesian engine is proprietary and access‑controlled** — available for pilot evaluation and enterprise licensing under outcome‑based pricing.
+
+Learn more at [github.com/arf-foundation](https://github.com/arf-foundation) and request access via petter2025us@outlook.com.