Add permanent saved voices + mobile-friendly premium UI

Adds two cleaned, normalised 20s reference voices to assets/ so no upload is needed. Quick-pick buttons and one-tap examples use them; the reference is preloaded by default. Shorter labels and CSS media queries fix cramped mobile rendering.

.gitattributes

@@ -43,3 +43,5 @@ training_data/19535.png filter=lfs diff=lfs merge=lfs -text
 training_data/19536.png filter=lfs diff=lfs merge=lfs -text
 training_data/19540.png filter=lfs diff=lfs merge=lfs -text
 training_data/19542.png filter=lfs diff=lfs merge=lfs -text
+assets/voice_1.wav filter=lfs diff=lfs merge=lfs -text
+assets/voice_2.wav filter=lfs diff=lfs merge=lfs -text

app.py

@@ -111,18 +111,13 @@ def _patch_numpy_compat() -> None:
         np.bool8 = np.bool_
 
 
-def _download_example_voice() -> Optional[str]:
-    """Fetch one sample reference voice so users can try the app instantly."""
-    try:
-        from huggingface_hub import hf_hub_download
-
-        return hf_hub_download(
-            repo_id="IndexTeam/IndexTTS-2-Demo",
-            repo_type="space",
-            filename="examples/voice_02.wav",
-        )
-    except Exception:
-        return None
+ASSETS_DIR = Path(__file__).resolve().parent / "assets"
+MY_VOICES = [
+    ("🎙️ My Voice 1", ASSETS_DIR / "voice_1.wav"),
+    ("🎙️ My Voice 2", ASSETS_DIR / "voice_2.wav"),
+]
+MY_VOICES = [(label, str(path)) for label, path in MY_VOICES if path.is_file()]
+DEFAULT_VOICE = MY_VOICES[0][1] if MY_VOICES else None
 
 
 def _warm_up() -> None:
@@ -148,9 +143,6 @@ def _warm_up() -> None:
 _warmup_thread = threading.Thread(target=_warm_up, daemon=True)
 _warmup_thread.start()
 
-EXAMPLE_VOICE = _download_example_voice()
-
-
 def _find_config(model_dir: Path) -> str:
     candidates = [
         model_dir / "config.yaml",
@@ -475,6 +467,16 @@ CUSTOM_CSS = """
 .quota-note {font-size: 0.78rem; color: #64748b; margin-top: 6px; line-height: 1.5;}
 .tips-card {font-size: 0.92rem; line-height: 1.7;}
 footer {display: none !important;}
+#foot {text-align: center; color: #64748b; margin-top: 20px; font-size: 0.85rem;}
+#foot span {opacity: .75;}
+@media (max-width: 768px) {
+  #hero {padding: 18px 16px; border-radius: 14px;}
+  #hero h1 {font-size: 1.35rem;}
+  #hero p {font-size: 0.85rem;}
+  .gradio-container {padding: 8px !important;}
+  .quota-row {flex-direction: column; gap: 1px;}
+  #generate-btn {height: 46px !important; font-size: 0.95rem !important;}
+}
 """
 
 HEADER_HTML = """
@@ -503,25 +505,33 @@ with gr.Blocks(title="Offex Voice Studio — AI Voice Cloning", theme=gr.themes.
     with gr.Row():
         with gr.Column(scale=5):
             reference_audio = gr.Audio(
-                label="🎧 Reference voice (5–15s of clean speech works best)",
+                label="Reference voice",
                 sources=["upload", "microphone"],
                 type="filepath",
+                value=DEFAULT_VOICE,
             )
-            with gr.Accordion("🎭 Emotion reference (optional — big quality boost)", open=False):
+            if MY_VOICES:
+                gr.Markdown("**⚡ Quick-pick — no upload needed:**")
+                with gr.Row():
+                    for label, path in MY_VOICES:
+                        gr.Button(label, size="sm").click(
+                            fn=lambda p=path: p, outputs=reference_audio
+                        )
+            with gr.Accordion("🎭 Emotion reference (optional — big boost)", open=False):
                 emotion_audio = gr.Audio(
-                    label="Emotion reference (same voice, desired emotion)",
+                    label="Emotion reference",
                     sources=["upload", "microphone"],
                     type="filepath",
                 )
         with gr.Column(scale=7):
             text = gr.Textbox(
-                label="📝 Text to speak (English)",
+                label="Text to speak",
                 placeholder="Type the text you want in the cloned voice...",
                 lines=5,
                 max_lines=12,
             )
             permission = gr.Checkbox(
-                label="✅ I have permission to use this reference voice",
+                label="✅ I have permission to use this voice",
                 value=False,
             )
             generate = gr.Button("✨ Generate Voice", variant="primary", size="lg", elem_id="generate-btn")
@@ -574,14 +584,16 @@ with gr.Blocks(title="Offex Voice Studio — AI Voice Cloning", theme=gr.themes.
     with gr.Accordion("💡 Tips for the most accurate clone", open=False):
         gr.Markdown(TIPS_MD)
 
-    if EXAMPLE_VOICE:
+    gr.HTML("<div id='foot'>Built with <span>❤️</span> in India · Powered by IndexTTS-2 on ZeroGPU</div>")
+
+    if MY_VOICES:
         gr.Examples(
             examples=[
-                [EXAMPLE_VOICE, "Hello there! This is my cloned voice, speaking naturally and clearly."],
-                [EXAMPLE_VOICE, "Welcome to Offex Voice Studio. Today is a wonderful day to create something amazing."],
+                [MY_VOICES[0][1], "Hello there! This is my cloned voice, speaking naturally and clearly."],
+                [MY_VOICES[-1][1], "Welcome to Offex Voice Studio. Today is a wonderful day to create something amazing."],
             ],
             inputs=[reference_audio, text],
-            label="Try it instantly with a sample voice",
+            label="One-tap examples with your saved voices",
         )
 
 

assets/voice_1.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:539b877c780d3c13645f8dad702269a94125121cc23819cc291dbe5daa720767
+size 1764044

assets/voice_2.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:acb7378e322eb12d8836281972e6f0c9ff1ea8ef1aee9d8dc2e670f8622f2c0d
+size 1764044