Update README.md

Author: jegly

README.md

@@ -4,7 +4,7 @@
 [![License](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](LICENSE)
 [![Upstream](https://img.shields.io/badge/upstream-google--ai--edge%2Fgallery-brightgreen)](https://github.com/google-ai-edge/gallery)
 
-**A security-hardened fork of [Google AI Edge Gallery](https://github.com/google-ai-edge/gallery) — with biometric lock, encrypted chat history, llama.cpp support, and GGUF model import.**
+**A security-hardened fork of [Google AI Edge Gallery](https://github.com/google-ai-edge/gallery) — with Unique Hybrid Features,biometric lock, encrypted chat history, llama.cpp support, and GGUF model import.**
 
 ## Disclaimer
 
@@ -33,6 +33,46 @@ Box is an Android app for running large language models entirely on-device. It i
 
 On the inference side, Box integrates llama.cpp alongside the upstream LiteRT runtime. This lets you sideload any GGUF model file and choose between CPU, GPU, or NPU acceleration per model — so you are not limited to the curated model list.
 
+# 🔒 Box: The Only Android App That Fuses Google's LiteRT with llama.cpp + Biometric Security
+
+**What makes Box unique?** While other on-device LLM apps force you to choose between Google's optimized LiteRT ecosystem (limited model selection) or the open GGUF ecosystem (limited hardware acceleration), **Box runs both side-by-side** — letting you import any GGUF model while keeping LiteRT's NPU acceleration for compatible models.
+
+## Why This Matters (And Why No One Else Does It)
+
+| Feature | Google AI Edge Gallery | llama.cpp-only apps (OfflineLLM) | **Box (This Project)** |
+|---------|----------------------|--------------------------------|------------------------|
+| LiteRT + NPU acceleration | ✅ | ❌ | ✅ |
+| Import any GGUF model | ❌ | ✅ | ✅ |
+| Encrypted chat history | ❌ | ❌ | ✅ |
+| Biometric app lock | ❌ | ❌ | ✅ |
+| Per-model accelerator choice | ❌ | ❌ | ✅ |
+| Hard offline mode (airgap) | ❌ | ❌ | ✅ |
+
+**The technical breakthrough:** Box doesn't just bundle both runtimes — it lets you choose **per-model** whether to run on CPU, GPU (via OpenCL/Vulkan), or NPU (via QNN delegate). No other Android app gives you this granular control.
+
+## The Hybrid Architecture Most Developers Think Is Impossible
+User's GGUF file → llama.cpp → CPU/GPU
+Google's .litertlm → LiteRT → NPU (Qualcomm/MediaTek)
+↓
+Same chat interface, same encrypted history
+
+
+Most developers assume you have to pick one inference engine. Box proves otherwise — and adds enterprise-grade security on top.
+
+## Built By Someone Who Already Did The Hard Part
+
+This isn't a theoretical project. I built **OfflineLLM** (pure llama.cpp app) first, then forked Google AI Edge Gallery to add llama.cpp support. The result: an app that inherits Google's polished UI and multimodal features (Ask Image, Audio Scribe, Agent Skills) while adding the open model flexibility that Google's curated allowlist prevents.
+
+## For Security-Conscious Users Running Sensitive Conversations
+
+- SQLCipher AES-256 encrypted Room database
+- Biometric re-authentication on every foreground
+- Hard offline switch — blocks all network traffic
+- Input sanitization before inference AND persistence
+- On-device security audit log
+
+**Bottom line:** If you want to run Qwen3.6,Llama 3, Mistral, or any GGUF model *with* the option of NPU acceleration when available — while keeping your conversations encrypted and offline — Box is currently the only Android app that delivers all of that in one package.
+
 ---
 
 ## What's different from upstream