Update README.md

README.md

@@ -1,13 +1,45 @@
 ---
-title: C2C Demo
-emoji: 📈
-colorFrom: pink
-colorTo: red
+title: Cache-to-Cache Communication Demo
+emoji: 🔗
+colorFrom: blue
+colorTo: blue
 sdk: gradio
-sdk_version: 5.49.1
+sdk_version: 5.9.1
 app_file: app.py
 pinned: false
 license: apache-2.0
+tags:
+- llm
+- cache-to-cache
+- model-communication
+- kv-cache
+short_description: Compare Single, Text-to-Text, and Cache-to-Cache inference
+thumbnail: >-
+  https://cdn-uploads.huggingface.co/production/uploads/6445fd9ba56444c355dcbcba/R5YOyw0aoBENYJs8Ugnbi.png
 ---
+# Cache-to-Cache Communication Demo
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+This Space demonstrates **Cache-to-Cache (C2C)** communication between Large Language Models, comparing three inference approaches side-by-side:
+
+1. **Single Model**: Standard inference with one model
+2. **Text-to-Text (T2T)**: Two-stage communication where Sharer model generates text → Receiver model processes text
+3. **Cache-to-Cache (C2C)**: Direct KV-Cache communication between Sharer and Receiver
+
+## What is Cache-to-Cache?
+
+It makes language models talk without words.
+
+Cache-to-Cache (C2C) lets multiple LLMs communicate directly through their KV-caches instead of text, transferring deep semantics without token-by-token generation. 
+
+The payoff: up to 10% higher accuracy, 3–5% gains over text-based communication, and 2× faster responses.
+
+## Citation
+
+```bibtex
+@article{fu2025c2c,
+    title={Cache-to-Cache: Direct Semantic Communication Between Large Language Models}, 
+    author={Tianyu Fu and Zihan Min and Hanling Zhang and Jichao Yan and Guohao Dai and Wanli Ouyang and Yu Wang},
+    journal={arXiv preprint arXiv:2510.03215},
+    year={2025},
+}
+```