Update README.md

README.md

@@ -1,29 +1,18 @@
----
-title: FugthDesign AI Backend
-emoji: 🤖
-colorFrom: purple
-colorTo: blue
-sdk: docker
-app_port: 7860
----
+# FugthDes Story Generator (Hugging Face Space)
 
-## FugthDesign AI Backend
+Lightweight CPU-based story generator using **TinyLlama GGUF** + **Flask API**.
 
-This Space provides the backend API for the floating AI assistant on the [fugthdesign.space](http://fugthdesign.space/) website.
+### 🚀 Features
+- Story continuation and feedback-based rewriting
+- Supports story memory (multi-part continuity)
+- CPU-friendly (TinyLlama Q4_K_M)
+- Connectable to GitHub front-end or any REST client
 
-### Configuration
-- **Hardware:** This space is configured to run on **CPU hardware**.
-- **Model:** `stablelm-zephyr-3b.Q3_K_S.gguf`
-
-### Performance Note
-⚠️ This Space is running on a free CPU and is intended for demonstration purposes only. Response times will be slow. For real-time performance, upgrade to a GPU hardware in the settings.
-
-### API Endpoint: `/chat`
-- **Method:** POST
-- **Body (JSON):**
+### 🧠 Endpoint
+POST `/generate`
 ```json
 {
-  "userInput": "Hello, how are you?",
-  "persona": "You are a helpful assistant.",
-  "localKnowledge": "Some extra context for the AI."
-}
+  "prompt": "Write chapter 1 of a fantasy story.",
+  "feedback": "Make it more emotional.",
+  "story_memory": "Previous chapters..."
+}