Update app.py

app.py

@@ -3,15 +3,27 @@ import threading
 
 import torch
 import spaces
+import spaces.zero
 
-from flask import Flask, jsonify, request, render_template_string
+from flask import Flask, request, jsonify, render_template_string
 from flask_cors import CORS
+
 from transformers import AutoTokenizer, AutoModelForCausalLM
 
 
 # ============================================================
 # WAAMU AI
-# Hugging Face GPU Space + Flask API
+# ============================================================
+
+APP_NAME = "WAAMU AI"
+MODEL_NAME = "Qwen/Qwen2.5-0.5B-Instruct"
+
+PORT = 7860
+GPU_DURATION = 120
+
+
+# ============================================================
+# Flask
 # ============================================================
 
 app = Flask(__name__)
@@ -27,131 +39,93 @@ CORS(
 
 
 # ============================================================
-# CONFIG
+# Model
+#
+# ZeroGPU expects the model to be placed on CUDA at module
+# level. The GPU decorator controls actual GPU execution.
 # ============================================================
 
-MODEL_NAME = os.getenv(
-    "MODEL_NAME",
-    "Qwen/Qwen2.5-0.5B-Instruct"
-)
+print("Starting WAAMU AI...", flush=True)
+print(f"Model: {MODEL_NAME}", flush=True)
 
-MAX_NEW_TOKENS = 256
-MAX_INPUT_CHARS = 12000
+print("Loading tokenizer...", flush=True)
 
+tokenizer = AutoTokenizer.from_pretrained(
+    MODEL_NAME
+)
 
-# ============================================================
-# SYSTEM PROMPT
-# ============================================================
+print("Loading model...", flush=True)
 
-SYSTEM_PROMPT = """
-You are WAAMU AI.
+model = AutoModelForCausalLM.from_pretrained(
+    MODEL_NAME,
+    torch_dtype=torch.float16
+)
 
-Your name is ALWAYS WAAMU AI.
+model.to("cuda")
+model.eval()
 
-IDENTITY RULES:
-- If the user asks your name, say: "My name is WAAMU AI."
-- If the user asks who you are, identify yourself as WAAMU AI.
-- If the user asks what AI or model you are, identify yourself as WAAMU AI.
-- Never claim that your name is Qwen.
-- Never replace your identity with the underlying model.
-- Qwen is only the underlying technology powering WAAMU AI.
+print("WAAMU AI model loaded.", flush=True)
 
-TEMPORARY INFORMATION:
-- Information supplied through the "info" field is temporary.
-- Use it only for the current request.
-- Never save it.
-- Never claim that you permanently remember it.
-- Do not create persistent memory.
 
-Answer naturally and helpfully.
-"""
+# Prevent multiple requests from simultaneously generating
+# with the same model.
+generation_lock = threading.Lock()
 
 
 # ============================================================
-# MODEL
+# WAAMU AI SYSTEM PROMPT
 # ============================================================
 
-tokenizer = None
-model = None
-
-model_lock = threading.Lock()
-
-
-def load_model():
-    global tokenizer
-    global model
-
-    if model is not None:
-        return
-
-    with model_lock:
+SYSTEM_PROMPT = """
+You are WAAMU AI.
 
-        if model is not None:
-            return
+Your public identity is ALWAYS WAAMU AI.
 
-        print("==========================================")
-        print("        STARTING WAAMU AI")
-        print("==========================================")
-        print("Model:", MODEL_NAME)
-        print("Loading tokenizer...")
+If the user asks:
+- What is your name?
+- Who are you?
+- What AI are you?
+- What model are you?
+- Are you Qwen?
+- What powers you?
 
-        tokenizer = AutoTokenizer.from_pretrained(
-            MODEL_NAME
-        )
+Answer that you are WAAMU AI.
 
-        print("Loading model...")
+Do not identify yourself as Qwen, Qwen2.5, Hugging Face,
+Transformers, or any other underlying model.
 
-        model = AutoModelForCausalLM.from_pretrained(
-            MODEL_NAME,
-            dtype=torch.float16
-        )
+You are a helpful general-purpose AI assistant.
 
-        model = model.to("cuda")
+The "info" supplied by the API is temporary context for the
+current request only.
 
-        model.eval()
+NEVER save, remember, persist, store, write, cache, or reuse
+the supplied info after the current request finishes.
 
-        print(
-            "WAAMU AI model loaded on:",
-            next(model.parameters()).device
-        )
+There is NO persistent memory.
 
-        print("==========================================")
+Be concise, helpful, and natural.
+"""
 
 
 # ============================================================
-# ACTUAL GPU FUNCTION
+# GPU GENERATION
 # ============================================================
-#
-# IMPORTANT:
-# This is a top-level @spaces.GPU function.
-# Hugging Face can detect it directly.
-#
-# ============================================================
-
-@spaces.GPU
-def waamu_generate(
-    message: str,
-    info: str = ""
-):
 
-    load_model()
-
-    message = str(
-        message or ""
-    ).strip()
-
-    info = str(
-        info or ""
-    ).strip()
+@spaces.GPU(duration=GPU_DURATION)
+def waamu_generate(message: str, info: str = ""):
+    """
+    Generate one WAAMU AI response.
 
+    info is temporary request context only.
+    It is never persisted.
+    """
 
-    if len(message) > MAX_INPUT_CHARS:
-        message = message[:MAX_INPUT_CHARS]
-
-
-    if len(info) > MAX_INPUT_CHARS:
-        info = info[:MAX_INPUT_CHARS]
+    message = str(message or "").strip()
+    info = str(info or "").strip()
 
+    if not message:
+        return "Please send a message."
 
     messages = [
         {
@@ -160,19 +134,20 @@ def waamu_generate(
         }
     ]
 
-
+    # Temporary context only.
     if info:
-
         messages.append(
             {
                 "role": "system",
-                "content":
-                    "Temporary information for this "
-                    "request only:\n\n" + info
+                "content": (
+                    "Temporary context supplied by the user "
+                    "for this request only:\n"
+                    f"{info}\n\n"
+                    "Do not save or remember this context."
+                )
             }
         )
 
-
     messages.append(
         {
             "role": "user",
@@ -180,1215 +155,549 @@ def waamu_generate(
         }
     )
 
-
     prompt = tokenizer.apply_chat_template(
         messages,
         tokenize=False,
         add_generation_prompt=True
     )
 
-
     inputs = tokenizer(
         prompt,
         return_tensors="pt"
     )
 
-
     inputs = {
         key: value.to("cuda")
         for key, value in inputs.items()
     }
 
+    with generation_lock:
+        with torch.inference_mode():
+            output = model.generate(
+                **inputs,
+                max_new_tokens=256,
+                do_sample=True,
+                temperature=0.7,
+                top_p=0.9,
+                repetition_penalty=1.05,
+                pad_token_id=tokenizer.eos_token_id
+            )
 
-    with torch.inference_mode():
-
-        output = model.generate(
-
-            **inputs,
-
-            max_new_tokens=MAX_NEW_TOKENS,
-
-            do_sample=True,
-
-            temperature=0.7,
-
-            top_p=0.9,
-
-            repetition_penalty=1.08,
-
-            pad_token_id=
-                tokenizer.eos_token_id
-        )
-
-
-    input_length = (
-        inputs["input_ids"].shape[1]
-    )
-
-
-    generated_tokens = output[
-        0,
-        input_length:
-    ]
-
+    generated_tokens = output[0][inputs["input_ids"].shape[1]:]
 
     reply = tokenizer.decode(
         generated_tokens,
         skip_special_tokens=True
     ).strip()
 
-
     if not reply:
-
-        reply = (
-            "I couldn't generate a response."
-        )
-
+        reply = "I couldn't generate a response."
 
     return reply
 
 
 # ============================================================
-# HOME PAGE
+# WEB UI
 # ============================================================
 
-HTML = r"""
+HTML = """
 <!DOCTYPE html>
-
-<html>
-
+<html lang="en">
 <head>
+    <meta charset="UTF-8">
+    <meta
+        name="viewport"
+        content="width=device-width, initial-scale=1.0"
+    >
 
-<meta charset="UTF-8">
-
-<meta
-name="viewport"
-content="width=device-width,initial-scale=1.0"
->
-
-<meta
-name="theme-color"
-content="#08030f"
->
-
-<title>WAAMU AI</title>
-
-
-<style>
-
-* {
-    box-sizing: border-box;
-}
-
-html,
-body {
-    margin: 0;
-    min-height: 100%;
-}
-
-body {
-
-    background:
-        radial-gradient(
-            circle at 15% 0%,
-            rgba(139,92,246,.20),
-            transparent 35%
-        ),
-        radial-gradient(
-            circle at 90% 100%,
-            rgba(191,0,255,.14),
-            transparent 35%
-        ),
-        #08030f;
-
-    color: #f7f0ff;
-
-    font-family:
-        Inter,
-        system-ui,
-        -apple-system,
-        BlinkMacSystemFont,
-        "Segoe UI",
-        sans-serif;
-}
-
-.app {
-
-    width: min(
-        1050px,
-        100%
-    );
-
-    min-height: 100vh;
-
-    margin: auto;
-
-    display: flex;
-
-    flex-direction: column;
-}
-
-
-/* HEADER */
-
-header {
-
-    height: 70px;
-
-    padding: 0 18px;
-
-    display: flex;
-
-    align-items: center;
-
-    justify-content: space-between;
-
-    position: sticky;
-
-    top: 0;
+    <title>WAAMU AI</title>
 
-    z-index: 10;
-
-    background:
-        rgba(8,3,15,.86);
-
-    backdrop-filter:
-        blur(18px);
-
-    border-bottom:
-        1px solid
-        rgba(190,100,255,.16);
-}
-
-.brand {
-
-    display: flex;
-
-    align-items: center;
-
-    gap: 11px;
-}
-
-.logo {
-
-    width: 42px;
-    height: 42px;
-
-    display: flex;
-
-    align-items: center;
-
-    justify-content: center;
-
-    border-radius: 13px;
-
-    font-weight: 900;
-
-    background:
-        linear-gradient(
-            135deg,
-            #7c3aed,
-            #bf00ff
-        );
-
-    box-shadow:
-        0 0 28px
-        rgba(191,0,255,.30);
-}
-
-.name {
-
-    font-size: 18px;
-
-    font-weight: 850;
-}
+    <style>
+        * {
+            box-sizing: border-box;
+        }
 
-.subtitle {
+        html,
+        body {
+            margin: 0;
+            padding: 0;
+            width: 100%;
+            min-height: 100%;
+            background: #07030d;
+            color: #ffffff;
+            font-family:
+                Inter,
+                system-ui,
+                -apple-system,
+                BlinkMacSystemFont,
+                "Segoe UI",
+                sans-serif;
+        }
 
-    margin-top: 2px;
+        body {
+            min-height: 100vh;
+            background:
+                radial-gradient(
+                    circle at top,
+                    rgba(139, 92, 246, 0.18),
+                    transparent 40%
+                ),
+                #07030d;
+        }
 
-    font-size: 11px;
+        .container {
+            width: min(900px, 94%);
+            margin: auto;
+            padding: 28px 0;
+        }
 
-    color: #a99bb8;
-}
+        .header {
+            text-align: center;
+            margin-bottom: 24px;
+        }
 
-.status {
+        .logo {
+            font-size: 34px;
+            font-weight: 900;
+            letter-spacing: -1px;
+            background: linear-gradient(
+                90deg,
+                #a855f7,
+                #d946ef,
+                #8b5cf6
+            );
+            -webkit-background-clip: text;
+            background-clip: text;
+            color: transparent;
+        }
 
-    font-size: 12px;
+        .subtitle {
+            margin-top: 7px;
+            color: #a99fb8;
+            font-size: 14px;
+        }
 
-    color: #b9a9c8;
-}
+        .card {
+            background: rgba(18, 9, 30, 0.88);
+            border: 1px solid rgba(168, 85, 247, 0.15);
+            border-radius: 20px;
+            padding: 18px;
+            box-shadow:
+                0 20px 70px rgba(0, 0, 0, 0.45),
+                0 0 45px rgba(139, 92, 246, 0.06);
+            backdrop-filter: blur(18px);
+        }
 
-.dot {
+        .info-label {
+            color: #b89cff;
+            font-size: 13px;
+            font-weight: 700;
+            margin-bottom: 7px;
+        }
 
-    display: inline-block;
+        .info {
+            width: 100%;
+            min-height: 70px;
+            resize: vertical;
+            border: 0;
+            outline: none;
+            border-radius: 14px;
+            padding: 13px;
+            background: #0c0713;
+            color: #ffffff;
+            font-size: 14px;
+            margin-bottom: 14px;
+        }
 
-    width: 7px;
-    height: 7px;
+        .message {
+            width: 100%;
+            min-height: 120px;
+            resize: vertical;
+            border: 0;
+            outline: none;
+            border-radius: 14px;
+            padding: 15px;
+            background: #0c0713;
+            color: #ffffff;
+            font-size: 15px;
+            margin-bottom: 14px;
+        }
 
-    margin-right: 5px;
+        textarea::placeholder {
+            color: #756a80;
+        }
 
-    border-radius: 50%;
+        .send {
+            width: 100%;
+            border: 0;
+            outline: none;
+            border-radius: 14px;
+            padding: 14px;
+            background: linear-gradient(
+                135deg,
+                #7c3aed,
+                #a855f7
+            );
+            color: white;
+            font-size: 15px;
+            font-weight: 800;
+            cursor: pointer;
+            transition: 0.2s ease;
+        }
 
-    background: #a855f7;
+        .send:hover {
+            transform: translateY(-1px);
+            filter: brightness(1.08);
+        }
 
-    box-shadow:
-        0 0 10px #a855f7;
-}
+        .send:disabled {
+            opacity: 0.55;
+            cursor: wait;
+            transform: none;
+        }
 
+        .response {
+            margin-top: 18px;
+            white-space: pre-wrap;
+            line-height: 1.6;
+            color: #eee8f5;
+            background: #0c0713;
+            border-radius: 14px;
+            padding: 15px;
+            min-height: 60px;
+            display: none;
+        }
 
-/* CHAT */
+        .temporary {
+            color: #8d8299;
+            font-size: 11px;
+            margin-top: -8px;
+            margin-bottom: 14px;
+        }
 
-.chat {
+        @media (max-width: 600px) {
+            .container {
+                width: 94%;
+                padding: 20px 0;
+            }
 
-    flex: 1;
+            .logo {
+                font-size: 29px;
+            }
 
-    padding:
-        20px
-        15px
-        285px;
-}
+            .card {
+                padding: 14px;
+                border-radius: 17px;
+            }
+        }
+    </style>
+</head>
 
-.welcome {
+<body>
 
-    text-align: center;
+<div class="container">
+
+    <div class="header">
+        <div class="logo">WAAMU AI</div>
+        <div class="subtitle">
+            Lightweight AI assistant
+        </div>
+    </div>
+
+    <div class="card">
+
+        <div class="info-label">
+            Temporary Info
+        </div>
+
+        <textarea
+            id="info"
+            class="info"
+            placeholder="Optional context for this request..."
+        ></textarea>
+
+        <div class="temporary">
+            ⚡ This information is used only for this request and is
+            never saved as memory.
+        </div>
+
+        <textarea
+            id="message"
+            class="message"
+            placeholder="Ask WAAMU AI anything..."
+        ></textarea>
+
+        <button
+            id="send"
+            class="send"
+            onclick="sendMessage()"
+        >
+            Send
+        </button>
+
+        <div
+            id="response"
+            class="response"
+        ></div>
+
+    </div>
 
-    padding:
-        70px
-        15px
-        35px;
-}
+</div>
 
-.welcome-logo {
 
-    width: 72px;
-    height: 72px;
+<script>
 
-    margin:
-        0 auto 18px;
+async function sendMessage() {
 
-    display: flex;
+    const message =
+        document.getElementById("message").value.trim();
 
-    align-items: center;
+    const info =
+        document.getElementById("info").value.trim();
 
-    justify-content: center;
+    const button =
+        document.getElementById("send");
 
-    border-radius: 23px;
+    const response =
+        document.getElementById("response");
 
-    font-size: 27px;
+    if (!message) {
+        response.style.display = "block";
+        response.textContent = "Please enter a message.";
+        return;
+    }
 
-    font-weight: 900;
+    button.disabled = true;
+    button.textContent = "Thinking...";
 
-    background:
-        linear-gradient(
-            135deg,
-            #7c3aed,
-            #bf00ff
-        );
+    response.style.display = "block";
+    response.textContent = "WAAMU AI is thinking...";
 
-    box-shadow:
-        0 0 45px
-        rgba(191,0,255,.30);
-}
+    try {
 
-.welcome h1 {
+        const result = await fetch(
+            "/chat",
+            {
+                method: "POST",
 
-    margin:
-        0 0 8px;
+                headers: {
+                    "Content-Type": "application/json"
+                },
 
-    font-size:
-        clamp(
-            28px,
-            7vw,
-            43px
+                body: JSON.stringify({
+                    message: message,
+                    info: info
+                })
+            }
         );
-}
-
-.welcome p {
-
-    margin: 0;
-
-    color: #a99bb8;
-
-    font-size: 14px;
-}
-
-
-/* MESSAGE */
-
-.message {
 
-    width:
-        min(
-            880px,
-            100%
-        );
+        const data = await result.json();
 
-    margin:
-        0 auto 18px;
+        if (data.success) {
+            response.textContent = data.reply;
+        } else {
+            response.textContent =
+                data.error || "Something went wrong.";
+        }
 
-    display: flex;
+    } catch (error) {
 
-    align-items: flex-start;
+        response.textContent =
+            "Unable to connect to WAAMU AI.";
 
-    gap: 10px;
-}
+    } finally {
 
-.message.user {
+        button.disabled = false;
+        button.textContent = "Send";
 
-    flex-direction: row-reverse;
+    }
 }
 
-.avatar {
-
-    width: 34px;
-
-    min-width: 34px;
-
-    height: 34px;
-
-    display: flex;
-
-    align-items: center;
-
-    justify-content: center;
+document
+    .getElementById("message")
+    .addEventListener("keydown", function(event) {
 
-    border-radius: 11px;
-
-    font-size: 10px;
-
-    font-weight: 850;
-
-    background: #181020;
-
-    border:
-        1px solid
-        rgba(190,100,255,.18);
-}
+        if (
+            event.key === "Enter" &&
+            !event.shiftKey
+        ) {
+            event.preventDefault();
+            sendMessage();
+        }
 
-.user .avatar {
+    });
 
-    background: #6d28d9;
-}
+</script>
 
-.bubble {
+</body>
+</html>
+"""
 
-    max-width:
-        min(
-            780px,
-            calc(100vw - 95px)
-        );
 
-    padding:
-        12px 15px;
+# ============================================================
+# ROUTES
+# ============================================================
 
-    border-radius: 16px;
+@app.route("/", methods=["GET"])
+def index():
+    return render_template_string(HTML)
 
-    font-size: 14px;
 
-    line-height: 1.6;
+@app.route("/health", methods=["GET"])
+def health():
+    return jsonify(
+        {
+            "success": True,
+            "name": APP_NAME,
+            "status": "online"
+        }
+    )
 
-    white-space: pre-wrap;
 
-    overflow-wrap: anywhere;
-}
+@app.route("/api", methods=["GET"])
+def api_info():
+    return jsonify(
+        {
+            "success": True,
+            "name": APP_NAME,
+            "model": APP_NAME,
+            "endpoints": {
+                "chat": "POST /chat",
+                "health": "GET /health"
+            },
+            "memory": False,
+            "temporary_info": True
+        }
+    )
 
-.assistant .bubble {
 
-    background: #171020;
+@app.route("/chat", methods=["POST"])
+def chat():
 
-    border:
-        1px solid
-        rgba(190,100,255,.17);
+    try:
 
-    border-top-left-radius: 5px;
-}
+        data = request.get_json(
+            silent=True
+        ) or {}
 
-.user .bubble {
+        message = data.get(
+            "message",
+            ""
+        )
 
-    background:
-        linear-gradient(
-            135deg,
-            #6d28d9,
-            #8b5cf6
-        );
+        info = data.get(
+            "info",
+            ""
+        )
 
-    border-top-right-radius: 5px;
-}
+        if not isinstance(message, str):
+            return jsonify(
+                {
+                    "success": False,
+                    "error": "message must be a string."
+                }
+            ), 400
 
+        if not isinstance(info, str):
+            info = str(info)
 
-/* COMPOSER */
+        message = message.strip()
+        info = info.strip()
 
-.composer-wrap {
+        if not message:
 
-    position: fixed;
+            return jsonify(
+                {
+                    "success": False,
+                    "error": "Missing message."
+                }
+            ), 400
 
-    left: 0;
-    right: 0;
-    bottom: 0;
+        reply = waamu_generate(
+            message,
+            info
+        )
 
-    z-index: 20;
+        # No database.
+        # No file.
+        # No session.
+        # No persistent memory.
 
-    padding:
-        12px
-        13px
-        16px;
+        return jsonify(
+            {
+                "success": True,
+                "name": APP_NAME,
+                "reply": reply
+            }
+        )
 
-    background:
-        linear-gradient(
-            to top,
-            #08030f 55%,
-            transparent
-        );
-}
+    except Exception as e:
 
-.composer {
+        print(
+            f"Chat error: {e}",
+            flush=True
+        )
 
-    width:
-        min(
-            900px,
-            100%
-        );
+        return jsonify(
+            {
+                "success": False,
+                "name": APP_NAME,
+                "error": str(e)
+            }
+        ), 500
 
-    margin: auto;
 
-    overflow: hidden;
+# ============================================================
+# ZERO-GPU STARTUP REGISTRATION
+# ============================================================
 
-    border-radius: 20px;
+def _zerogpu_startup():
 
-    background:
-        rgba(16,7,28,.97);
+    try:
 
-    border:
-        1px solid
-        rgba(190,100,255,.18);
+        from spaces.zero import startup
 
-    box-shadow:
-        0 15px 60px
-        rgba(0,0,0,.50);
-}
+        startup()
 
-.info-header {
+        print(
+            "ZeroGPU startup registration sent.",
+            flush=True
+        )
 
-    display: flex;
+    except ImportError:
 
-    justify-content: space-between;
+        print(
+            "spaces.zero startup unavailable.",
+            flush=True
+        )
 
-    padding:
-        9px 13px;
+    except Exception as e:
 
-    border-bottom:
-        1px solid
-        rgba(190,100,255,.13);
+        print(
+            f"ZeroGPU startup registration failed: {e}",
+            flush=True
+        )
 
-    color: #b9a9c8;
 
-    font-size: 12px;
-}
+# ============================================================
+# START
+# ============================================================
 
-.not-saved {
+if __name__ == "__main__":
 
-    opacity: .55;
-}
+    # IMPORTANT:
+    # This must run AFTER @spaces.GPU has registered
+    # waamu_generate().
+    _zerogpu_startup()
 
-textarea {
-
-    width: 100%;
-
-    display: block;
-
-    resize: none;
-
-    outline: none;
-
-    border: 0;
-
-    background: transparent;
-
-    color: #f7f0ff;
-
-    font-family: inherit;
-
-    font-size: 14px;
-
-    padding:
-        11px 13px;
-}
-
-textarea::placeholder {
-
-    color: #75667f;
-}
-
-#info {
-
-    min-height: 52px;
-
-    max-height: 100px;
-}
-
-#message {
-
-    min-height: 55px;
-
-    max-height: 160px;
-
-    border-top:
-        1px solid
-        rgba(190,100,255,.10);
-}
-
-.send-row {
-
-    display: flex;
-
-    align-items: center;
-
-    justify-content: space-between;
-
-    padding:
-        7px
-        10px
-        10px
-        13px;
-}
-
-.hint {
-
-    color: #71637b;
-
-    font-size: 11px;
-}
-
-button {
-
-    border: 0;
-
-    outline: 0;
-
-    cursor: pointer;
-
-    padding:
-        10px 18px;
-
-    border-radius: 12px;
-
-    color: white;
-
-    font-weight: 750;
-
-    background:
-        linear-gradient(
-            135deg,
-            #7c3aed,
-            #bf00ff
-        );
-
-    box-shadow:
-        0 0 20px
-        rgba(191,0,255,.20);
-}
-
-button:disabled {
-
-    opacity: .5;
-
-    cursor: not-allowed;
-}
-
-
-/* MOBILE */
-
-@media(max-width:600px) {
-
-    header {
-
-        height: 62px;
-
-        padding:
-            0 13px;
-    }
-
-    .logo {
-
-        width: 38px;
-        height: 38px;
-    }
-
-    .name {
-
-        font-size: 16px;
-    }
-
-    .status {
-
-        display: none;
-    }
-
-    .chat {
-
-        padding:
-            10px
-            9px
-            285px;
-    }
-
-    .welcome {
-
-        padding:
-            55px
-            10px
-            30px;
-    }
-
-    .bubble {
-
-        font-size: 13.5px;
-    }
-
-    .hint {
-
-        display: none;
-    }
-
-    .composer-wrap {
-
-        padding:
-            8px
-            8px
-            10px;
-    }
-}
-
-</style>
-
-</head>
-
-
-<body>
-
-<div class="app">
-
-
-<header>
-
-<div class="brand">
-
-<div class="logo">
-W
-</div>
-
-<div>
-
-<div class="name">
-WAAMU AI
-</div>
-
-<div class="subtitle">
-Local model • Temporary context
-</div>
-
-</div>
-
-</div>
-
-
-<div class="status">
-
-<span class="dot"></span>
-
-Online
-
-</div>
-
-</header>
-
-
-<main
-id="chat"
-class="chat"
->
-
-<section
-id="welcome"
-class="welcome"
->
-
-<div class="welcome-logo">
-W
-</div>
-
-<h1>
-WAAMU AI
-</h1>
-
-<p>
-Ask anything. Add temporary information when needed.
-</p>
-
-</section>
-
-</main>
-
-
-<div class="composer-wrap">
-
-<div class="composer">
-
-
-<div class="info-header">
-
-<span>
-🧠 Temporary Info
-</span>
-
-<span class="not-saved">
-Not saved
-</span>
-
-</div>
-
-
-<textarea
-id="info"
-placeholder="Optional information for this request only..."
-></textarea>
-
-
-<textarea
-id="message"
-placeholder="Message WAAMU AI..."
-></textarea>
-
-
-<div class="send-row">
-
-<div class="hint">
-Enter to send • Shift+Enter for new line
-</div>
-
-<button
-id="send"
-onclick="sendMessage()"
->
-Send
-</button>
-
-</div>
-
-
-</div>
-
-</div>
-
-
-</div>
-
-
-<script>
-
-const chat =
-document.getElementById("chat");
-
-const messageInput =
-document.getElementById("message");
-
-const infoInput =
-document.getElementById("info");
-
-const sendButton =
-document.getElementById("send");
-
-
-function addMessage(
-    text,
-    type
-) {
-
-    const message =
-        document.createElement("div");
-
-    message.className =
-        "message " + type;
-
-
-    const avatar =
-        document.createElement("div");
-
-    avatar.className =
-        "avatar";
-
-    avatar.textContent =
-        type === "user"
-            ? "YOU"
-            : "W";
-
-
-    const bubble =
-        document.createElement("div");
-
-    bubble.className =
-        "bubble";
-
-    bubble.textContent =
-        text;
-
-
-    message.appendChild(
-        avatar
-    );
-
-    message.appendChild(
-        bubble
-    );
-
-    chat.appendChild(
-        message
-    );
-
-
-    window.scrollTo({
-        top:
-            document.body.scrollHeight,
-
-        behavior:
-            "smooth"
-    });
-
-
-    return bubble;
-}
-
-
-async function sendMessage() {
-
-    const message =
-        messageInput.value.trim();
-
-    const info =
-        infoInput.value.trim();
-
-
-    if (!message) {
-        return;
-    }
-
-
-    const welcome =
-        document.getElementById(
-            "welcome"
-        );
-
-    if (welcome) {
-        welcome.remove();
-    }
-
-
-    addMessage(
-        message,
-        "user"
-    );
-
-
-    messageInput.value = "";
-
-    sendButton.disabled = true;
-
-    sendButton.textContent =
-        "Thinking...";
-
-
-    const thinking =
-        addMessage(
-            "WAAMU AI is thinking...",
-            "assistant"
-        );
-
-
-    try {
-
-        const response =
-            await fetch(
-                "/chat",
-                {
-                    method: "POST",
-
-                    headers: {
-                        "Content-Type":
-                            "application/json"
-                    },
-
-                    body:
-                        JSON.stringify({
-                            message:
-                                message,
-
-                            info:
-                                info
-                        })
-                }
-            );
-
-
-        const data =
-            await response.json();
-
-
-        if (!response.ok) {
-
-            throw new Error(
-                data.error ||
-                "Request failed."
-            );
-        }
-
-
-        thinking.textContent =
-            data.reply ||
-            "No response.";
-
-
-    } catch (error) {
-
-        thinking.textContent =
-            "Error: " +
-            error.message;
-
-    } finally {
-
-        sendButton.disabled =
-            false;
-
-        sendButton.textContent =
-            "Send";
-
-        messageInput.focus();
-    }
-}
-
-
-messageInput.addEventListener(
-    "keydown",
-    function(event) {
-
-        if (
-            event.key === "Enter" &&
-            !event.shiftKey
-        ) {
-
-            event.preventDefault();
-
-            sendMessage();
-        }
-    }
-);
-
-</script>
-
-</body>
-
-</html>
-"""
-
-
-# ============================================================
-# FLASK ROUTES
-# ============================================================
-
-@app.route("/")
-def home():
-
-    return render_template_string(
-        HTML
-    )
-
-
-@app.route(
-    "/chat",
-    methods=["POST"]
-)
-def chat():
-
-    data = request.get_json(
-        silent=True
-    )
-
-
-    if not isinstance(
-        data,
-        dict
-    ):
-
-        return jsonify({
-            "success": False,
-            "error":
-                "JSON body required."
-        }), 400
-
-
-    message = data.get(
-        "message"
-    )
-
-
-    if message is None:
-
-        return jsonify({
-            "success": False,
-            "error":
-                "Missing message."
-        }), 400
-
-
-    message = str(
-        message
-    ).strip()
-
-
-    if not message:
-
-        return jsonify({
-            "success": False,
-            "error":
-                "Message cannot be empty."
-        }), 400
-
-
-    info = data.get(
-        "info",
-        ""
-    )
-
-    info = str(
-        info or ""
-    ).strip()
-
-
-    try:
-
-        reply = waamu_generate(
-            message,
-            info
-        )
-
-
-        return jsonify({
-
-            "success":
-                True,
-
-            "name":
-                "WAAMU AI",
-
-            "reply":
-                reply
-
-        })
-
-
-    except Exception as error:
-
-        print(
-            "WAAMU AI error:",
-            repr(error)
-        )
-
-
-        return jsonify({
-
-            "success":
-                False,
-
-            "name":
-                "WAAMU AI",
-
-            "error":
-                "AI generation failed."
-
-        }), 500
-
-
-# ============================================================
-# HEALTH
-# ============================================================
-
-@app.route("/health")
-def health():
-
-    return jsonify({
-
-        "status":
-            "online",
-
-        "name":
-            "WAAMU AI",
-
-        "model":
-            MODEL_NAME,
-
-        "memory":
-            False,
-
-        "cors":
-            True
-
-    })
-
-
-# ============================================================
-# API INFORMATION
-# ============================================================
-
-@app.route("/api")
-def api_info():
-
-    return jsonify({
-
-        "name":
-            "WAAMU AI",
-
-        "model":
-            MODEL_NAME,
-
-        "memory":
-            False,
-
-        "cors":
-            True,
-
-        "endpoints": {
-
-            "chat":
-                "POST /chat",
-
-            "home":
-                "GET /",
-
-            "health":
-                "GET /health"
-
-        }
-
-    })
-
-
-# ============================================================
-# START FLASK
-# ============================================================
-
-if __name__ == "__main__":
-
-    print(
-        "Starting WAAMU AI Flask API..."
-    )
+    print(
+        f"Starting {APP_NAME} Flask API...",
+        flush=True
+    )
 
     app.run(
         host="0.0.0.0",
-        port=int(
-            os.getenv(
-                "PORT",
-                "7860"
-            )
-        ),
+        port=PORT,
+        debug=False,
         threaded=True
     )