Update app.py

app.py

@@ -407,13 +407,15 @@ with gr.Blocks(theme=theme, css=light_blue_glass_css, title="Falconz Unified App
 
     gr.Markdown(""" # 🔐 Falconz - RedTeamers
 
-            ### 🛡️ Unified AI Security for Multi-Model & Agentic Systems
-            Falconz is an MCP-powered Gradio platform that safeguards LLM and agentic applications through real-time jailbreak and prompt-injection detection across OpenAI, Gemini, Mistral, Phi, and more.
-            It includes an inbuilt library of the latest Top 10 jailbreak templates (Inspired by O.S.W.A.P) that users can customize, modify, and deploy for controlled testing and red-teaming workflows.
-            Falconz supports both prompt template modification and threat detection—letting users experiment safely while continuously monitoring model behavior.
-            A lightweight safeguard model provides fast, on-device-friendly risk screening for quick evaluation cycles.
-            Plug-and-play with MCP to secure your AI stack and access live analytics in a single, streamlined interface.
-            """)
+                  ### 🛡️ Unified AI Security for Multi-Model & Agentic Systems
+                  Falconz is a Gradio MCP-powered platform designed to safeguard LLM and agentic applications through real-time jailbreak and prompt-injection detection across OpenAI, Gemini, Mistral, Phi, and more. Built on Gradio's intuitive interface integrated with MCP (Model Context Protocol), it enables seamless, plug-and-play security workflows.
+
+                  At its core, Falconz leverages **Anthropic's Claude models**—renowned for their robustness and resistance to prompt injections—as the primary engine for advanced threat detection in both text and image inputs. Claude's architecture makes it uniquely effective at identifying and mitigating subtle jailbreaks, vanilla injections, hidden templates, and adversarial prompts, providing unparalleled reliability for red-teaming tasks.
+
+                  The app includes an inbuilt library of the latest Top 10 jailbreak templates (inspired by OWASP), allowing users to customize, modify, and deploy them for controlled testing and ethical red-teaming workflows. It supports both prompt template modification and real-time threat detection, enabling safe experimentation while continuously monitoring model behavior. A lightweight safeguard model ensures fast, on-device-friendly risk screening for quick evaluation cycles.
+
+                  Integrate Falconz with MCP to secure your entire AI stack, access live analytics, and maintain ethical AI practices in a single, streamlined interface.
+                  """)
 
 
     with gr.Tabs():