Upload app.py with huggingface_hub

app.py

@@ -486,6 +486,12 @@ pretrained on 100M+ histopathology patches). A FiLM-based stain embedding allows
 - **Classifier-free guidance (CFG)**: 10% class dropout and 10% UNI dropout during training
   enables tunable generation strength at inference time.
 
+### Links
+
+- **Paper**: [arXiv:2603.12716](https://arxiv.org/abs/2603.12716)
+- **Code**: [github.com/facevoid/UNIStainNet](https://github.com/facevoid/UNIStainNet)
+- **Project Page**: [facevoid.github.io/UNIStainNet](https://facevoid.github.io/UNIStainNet/)
+
 ### Disclaimer
 
 This is a **research tool** for exploratory analysis. It is not intended for clinical diagnosis
@@ -496,7 +502,10 @@ and has not undergone regulatory validation. Generated stains should not be used
     # ── Footer ───────────────────────────────────────────────────
     gr.HTML("""
     <p style="text-align:center; padding:1rem; color:#999; font-size:0.8rem;">
-        UNIStainNet | Built with Gradio
+        UNIStainNet |
+        <a href="https://arxiv.org/abs/2603.12716" style="color:#888; text-decoration:none;">arXiv</a> |
+        <a href="https://github.com/facevoid/UNIStainNet" style="color:#888; text-decoration:none;">GitHub</a> |
+        <a href="https://facevoid.github.io/UNIStainNet/" style="color:#888; text-decoration:none;">Project Page</a>
     </p>
     """)