Update README.md

README.md

@@ -1,21 +1,38 @@
-# Universal Code Generator (Hugging Face Inference) — Gradio
+---
+title: Universal Code Generator
+emoji: 🤖
+colorFrom: blue
+colorTo: purple
+sdk: gradio
+sdk_version: "4.0.0"
+app_file: app.py
+pinned: false
+---
 
-This Space uses a Hugging Face Inference API model to generate, explain, fix and convert code.
+# 🔧 Universal Code Generator  
+A multi-language AI code generator powered by **Hugging Face Inference API**.
 
-## How to configure
+## 🚀 Features
+- Generate code in **70+ programming languages**
+- Explain code
+- Fix code
+- Convert code between languages
+- Runs entirely through HF Inference API (NO torch needed)
+- UI built with Gradio
 
-1. Add a secret in your Space settings:
-   - Name: `HF_TOKEN` (or `HUGGINGFACEHUB_API_TOKEN`)
-   - Value: your Hugging Face Inference API token
+## 🧠 Supported Models
+Recommended code models:
+- `Qwen/Qwen2.5-Coder-7B-Instruct`
+- `bigcode/starcoder2-3b`
+- `codellama/CodeLlama-7b-hf`
+- `microsoft/Phi-3-mini-4k-instruct`
+- `deepseek-ai/deepseek-coder-6.7b-instruct`
 
-2. Optional: change the default model in the UI. Recommended code models:
-   - `bigcode/starcoder2-3b`
-   - `codellama/CodeLlama-7b-hf`
-   - `Qwen/Qwen2.5-Coder-7B-Instruct`
+## 🔑 Usage
+Paste your **Hugging Face API Token** in the UI, select a model, and start generating.
 
-3. The app is built with Gradio. The Space will automatically start the app.
-
-## Notes
-- This app does **not** execute generated code. Always review code before running.
-- For larger models prefer using the Inference API rather than downloading big models into the Space.
+## ▶️ Run
+This Space uses:
 
+```sh
+python app.py