diff --git a/__pycache__/orchestrator.cpython-313.pyc b/__pycache__/orchestrator.cpython-313.pyc
index b604ee0..b0bb017 100644
Binary files a/__pycache__/orchestrator.cpython-313.pyc and b/__pycache__/orchestrator.cpython-313.pyc differ
diff --git a/config/best_config.json b/config/best_config.json
new file mode 100644
index 0000000..b6bbe12
--- /dev/null
+++ b/config/best_config.json
@@ -0,0 +1,68 @@
+{
+  "model_type": "xgboost",
+  "features": {
+    "use_price_position": true,
+    "use_momentum": true,
+    "use_volatility": true,
+    "use_volume": true,
+    "use_cycle": true,
+    "use_pca": false,
+    "pca_variance": 0.95,
+    "use_scaler": true
+  },
+  "target": {
+    "type": "regression",
+    "forward_periods_1h": [
+      168,
+      720,
+      2160
+    ],
+    "forward_periods_4h": [
+      42,
+      180,
+      540
+    ],
+    "weights": [
+      0.2,
+      0.3,
+      0.5
+    ],
+    "score_range": [
+      0,
+      100
+    ]
+  },
+  "hyperparameters": {
+    "learning_rate": 0.01,
+    "max_depth": 4,
+    "n_estimators": 300,
+    "subsample": 0.8,
+    "colsample_bytree": 0.8,
+    "min_child_weight": 20,
+    "gamma": 0.3,
+    "reg_alpha": 0.5,
+    "reg_lambda": 3.0,
+    "lstm_hidden_size": 128,
+    "lstm_num_layers": 2,
+    "lstm_dropout": 0.3,
+    "lstm_epochs": 100,
+    "lstm_batch_size": 64,
+    "lstm_sequence_length": 30,
+    "lstm_patience": 10
+  },
+  "strategy": {
+    "strong_buy_threshold": 65,
+    "good_buy_threshold": 55,
+    "poor_threshold": 35
+  },
+  "training": {
+    "rolling_window": true,
+    "rolling_train_size": 2500,
+    "rolling_test_size": 300,
+    "walk_forward_windows": 5,
+    "train_pct": 0.7,
+    "validation_pct": 0.15,
+    "test_pct": 0.15
+  },
+  "timeframe": "4h"
+}
\ No newline at end of file
diff --git a/config/current_config.json b/config/current_config.json
index 9ac6e87..b6bbe12 100644
--- a/config/current_config.json
+++ b/config/current_config.json
@@ -1,92 +1,68 @@
 {
-  "model_type": "hybrid",
+  "model_type": "xgboost",
   "features": {
-    "technical_indicators": [
-      "RSI_14",
-      "RSI_7",
-      "MACD_line",
-      "MACD_signal",
-      "MACD_hist",
-      "BB_upper",
-      "BB_lower",
-      "BB_width",
-      "ATR_14",
-      "SMA_20",
-      "SMA_50",
-      "EMA_10",
-      "EMA_20",
-      "OBV",
-      "stoch_k",
-      "stoch_d",
-      "williams_r",
-      "CCI_20",
-      "ROC_10"
-    ],
-    "lookback_periods": [
-      3,
-      5,
-      10,
-      20
-    ],
-    "use_volume_features": true,
-    "use_volatility_features": true,
-    "use_candle_patterns": false,
-    "use_lag_features": true,
-    "lag_periods": [
-      1,
-      2,
-      3,
-      5
-    ],
-    "use_pca": true,
+    "use_price_position": true,
+    "use_momentum": true,
+    "use_volatility": true,
+    "use_volume": true,
+    "use_cycle": true,
+    "use_pca": false,
     "pca_variance": 0.95,
     "use_scaler": true
   },
   "target": {
-    "type": "classification",
-    "direction": "both",
-    "horizon_candles": 4,
-    "threshold_pct": 1.0
+    "type": "regression",
+    "forward_periods_1h": [
+      168,
+      720,
+      2160
+    ],
+    "forward_periods_4h": [
+      42,
+      180,
+      540
+    ],
+    "weights": [
+      0.2,
+      0.3,
+      0.5
+    ],
+    "score_range": [
+      0,
+      100
+    ]
   },
   "hyperparameters": {
-    "learning_rate": 0.001,
-    "max_depth": 5,
+    "learning_rate": 0.01,
+    "max_depth": 4,
     "n_estimators": 300,
     "subsample": 0.8,
     "colsample_bytree": 0.8,
-    "min_child_weight": 5,
+    "min_child_weight": 20,
     "gamma": 0.3,
-    "reg_alpha": 0.1,
-    "reg_lambda": 5.0,
+    "reg_alpha": 0.5,
+    "reg_lambda": 3.0,
     "lstm_hidden_size": 128,
     "lstm_num_layers": 2,
     "lstm_dropout": 0.3,
     "lstm_epochs": 100,
     "lstm_batch_size": 64,
-    "lstm_sequence_length": 20,
+    "lstm_sequence_length": 30,
     "lstm_patience": 10
   },
   "strategy": {
-    "entry_threshold": 0.65,
-    "exit_type": "trailing_stop",
-    "stop_loss_pct": 2.0,
-    "take_profit_pct": 4.0,
-    "trailing_stop_pct": 1.5,
-    "position_sizing": "confidence_scaled",
-    "max_position_pct": 100,
-    "min_confidence_to_trade": 0.5,
-    "dynamic_sl_tp": true,
-    "atr_sl_multiplier": 1.2,
-    "atr_tp_multiplier": 3.0
+    "strong_buy_threshold": 65,
+    "good_buy_threshold": 55,
+    "poor_threshold": 35
   },
   "training": {
+    "rolling_window": true,
+    "rolling_train_size": 2500,
+    "rolling_test_size": 300,
     "walk_forward_windows": 5,
     "train_pct": 0.7,
     "validation_pct": 0.15,
-    "test_pct": 0.15,
-    "rolling_window": true,
-    "rolling_train_size": 3000,
-    "rolling_test_size": 200
+    "test_pct": 0.15
   },
   "timeframe": "4h"
 }
\ No newline at end of file
diff --git a/config/llm_settings.json b/config/llm_settings.json
new file mode 100644
index 0000000..2e532c5
--- /dev/null
+++ b/config/llm_settings.json
@@ -0,0 +1,21 @@
+{
+  "provider": "ollama",
+  "model": "qwen3.5:27b",
+  "providers": {
+    "ollama": {
+      "base_url": "http://100.100.242.21:11434"
+    },
+    "lmstudio": {
+      "base_url": "http://100.100.242.21:1234"
+    },
+    "openai": {
+      "api_key": ""
+    },
+    "anthropic": {
+      "api_key": ""
+    },
+    "openrouter": {
+      "api_key": ""
+    }
+  }
+}
diff --git a/dashboard/server.py b/dashboard/server.py
index 72aa0e8..1df07ab 100644
--- a/dashboard/server.py
+++ b/dashboard/server.py
@@ -9,6 +9,7 @@ import os
 import sys
 import threading
 
+import requests
 from fastapi import FastAPI
 from fastapi.responses import FileResponse, HTMLResponse, JSONResponse
 from pydantic import BaseModel
@@ -23,6 +24,7 @@ app = FastAPI(title="BTC Accumulation Signal Optimizer")
 CONFIG_DIR = os.path.join(BASE_DIR, "config")
 RESULTS_DIR = os.path.join(BASE_DIR, "results")
 ITERATIONS_LOG = os.path.join(RESULTS_DIR, "iterations.jsonl")
+LLM_SETTINGS_PATH = os.path.join(CONFIG_DIR, "llm_settings.json")
 
 _opt_thread = None
 
@@ -31,6 +33,64 @@ class ConfigUpdate(BaseModel):
     config: dict
 
 
+class LLMSettingsUpdate(BaseModel):
+    provider: str
+    model: str
+    providers: dict
+
+
+class TestConnectionRequest(BaseModel):
+    provider: str
+    providers: dict
+
+
+class FetchModelsRequest(BaseModel):
+    provider: str
+    providers: dict
+
+
+def _load_llm_settings():
+    if os.path.exists(LLM_SETTINGS_PATH):
+        with open(LLM_SETTINGS_PATH) as f:
+            return json.load(f)
+    return {
+        "provider": "ollama",
+        "model": "qwen3.5:27b",
+        "providers": {
+            "ollama": {"base_url": "http://100.100.242.21:11434"},
+            "lmstudio": {"base_url": "http://100.100.242.21:1234"},
+            "openai": {"api_key": ""},
+            "anthropic": {"api_key": ""},
+            "openrouter": {"api_key": ""},
+        },
+    }
+
+
+def _mask_api_key(key):
+    if not key or len(key) < 8:
+        return ""
+    return "••••••••" + key[-4:]
+
+
+def _safe_settings(settings):
+    """Return settings with API keys masked."""
+    out = json.loads(json.dumps(settings))
+    for name, cfg in out.get("providers", {}).items():
+        if "api_key" in cfg:
+            cfg["api_key"] = _mask_api_key(cfg["api_key"])
+    return out
+
+
+def _merge_api_keys(new_providers, existing_providers):
+    """Preserve existing API keys when the incoming value is masked."""
+    for name, cfg in new_providers.items():
+        if "api_key" in cfg:
+            masked = cfg["api_key"]
+            if masked.startswith("••••") or masked == "":
+                existing_key = existing_providers.get(name, {}).get("api_key", "")
+                cfg["api_key"] = existing_key
+
+
 @app.get("/api/status")
 def api_status():
     return orchestrator.get_status()
@@ -91,7 +151,11 @@ def api_best():
     with open(best_path) as f:
         config = json.load(f)
     iterations = orchestrator.load_iteration_history()
-    best_iter = max(iterations, key=lambda x: x.get("cost_improvement", 0)) if iterations else {}
+    best_iter = (
+        max(iterations, key=lambda x: x.get("cost_improvement", 0))
+        if iterations
+        else {}
+    )
     return {"config": config, "best_iteration": best_iter}
 
 
@@ -110,32 +174,155 @@ def api_download_best_config():
     return JSONResponse({"error": "No best config yet"}, status_code=404)
 
 
-DASHBOARD_HTML = """<!DOCTYPE html>
-<html lang="en">
-<head>
-<meta charset="UTF-8">
-<meta name="viewport" content="width=device-width, initial-scale=1.0">
-<title>BTC Accumulation Signal Optimizer</title>
-<link rel="preconnect" href="https://fonts.googleapis.com">
-<link href="https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700&display=swap" rel="stylesheet">
-<script src="https://cdn.jsdelivr.net/npm/chart.js@4.4.4/dist/chart.umd.min.js"></script>
-<style>
+# ── Settings API ──────────────────────────────────────────────────────────
+
+@app.get("/api/settings")
+def api_get_settings():
+    settings = _load_llm_settings()
+    return _safe_settings(settings)
+
+
+@app.post("/api/settings")
+def api_save_settings(body: LLMSettingsUpdate):
+    existing = _load_llm_settings()
+    new_settings = {
+        "provider": body.provider,
+        "model": body.model,
+        "providers": body.providers,
+    }
+    _merge_api_keys(new_settings["providers"], existing.get("providers", {}))
+    with open(LLM_SETTINGS_PATH, "w") as f:
+        json.dump(new_settings, f, indent=2)
+    return {"ok": True, "message": "Settings saved"}
+
+
+@app.post("/api/settings/test")
+def api_test_connection(body: TestConnectionRequest):
+    """Test connection to a provider and return available models."""
+    existing = _load_llm_settings()
+    providers = json.loads(json.dumps(body.providers))
+    _merge_api_keys(providers, existing.get("providers", {}))
+    provider = body.provider
+
+    try:
+        models = _fetch_models(provider, providers)
+        return {"ok": True, "models": models, "message": f"Connected — {len(models)} model(s) found"}
+    except requests.exceptions.ConnectionError:
+        return JSONResponse(
+            {"ok": False, "error": "Connection refused — is the server running?"},
+            status_code=502,
+        )
+    except requests.exceptions.Timeout:
+        return JSONResponse(
+            {"ok": False, "error": "Connection timed out"},
+            status_code=504,
+        )
+    except Exception as e:
+        return JSONResponse(
+            {"ok": False, "error": str(e)},
+            status_code=500,
+        )
+
+
+@app.post("/api/settings/models")
+def api_fetch_models(body: FetchModelsRequest):
+    """Fetch available models for a provider (proxied to avoid CORS)."""
+    existing = _load_llm_settings()
+    providers = json.loads(json.dumps(body.providers))
+    _merge_api_keys(providers, existing.get("providers", {}))
+
+    try:
+        models = _fetch_models(body.provider, providers)
+        return {"ok": True, "models": models}
+    except Exception as e:
+        return JSONResponse({"ok": False, "error": str(e)}, status_code=500)
+
+
+def _fetch_models(provider, providers):
+    """Fetch model list from a provider. Returns list of {id, name}."""
+    cfg = providers.get(provider, {})
+
+    if provider == "ollama":
+        base_url = cfg.get("base_url", "http://100.100.242.21:11434")
+        resp = requests.get(f"{base_url}/api/tags", timeout=10)
+        resp.raise_for_status()
+        data = resp.json()
+        return [{"id": m["name"], "name": m["name"]} for m in data.get("models", [])]
+
+    elif provider == "lmstudio":
+        base_url = cfg.get("base_url", "http://100.100.242.21:1234")
+        resp = requests.get(f"{base_url}/v1/models", timeout=10)
+        resp.raise_for_status()
+        data = resp.json()
+        return [{"id": m["id"], "name": m["id"]} for m in data.get("data", [])]
+
+    elif provider == "openai":
+        api_key = cfg.get("api_key", "")
+        if not api_key:
+            raise ValueError("OpenAI API key is required")
+        resp = requests.get(
+            "https://api.openai.com/v1/models",
+            headers={"Authorization": f"Bearer {api_key}"},
+            timeout=15,
+        )
+        resp.raise_for_status()
+        data = resp.json()
+        models = [m for m in data.get("data", []) if m["id"].startswith("gpt-") or "chat" in m.get("id", "")]
+        models.sort(key=lambda m: m["id"])
+        return [{"id": m["id"], "name": m["id"]} for m in models]
+
+    elif provider == "anthropic":
+        api_key = cfg.get("api_key", "")
+        if not api_key:
+            raise ValueError("Anthropic API key is required")
+        resp = requests.get(
+            "https://api.anthropic.com/v1/models",
+            headers={
+                "x-api-key": api_key,
+                "anthropic-version": "2023-06-01",
+            },
+            timeout=15,
+        )
+        resp.raise_for_status()
+        data = resp.json()
+        return [{"id": m["id"], "name": m.get("display_name", m["id"])} for m in data.get("data", [])]
+
+    elif provider == "openrouter":
+        resp = requests.get("https://openrouter.ai/api/v1/models", timeout=15)
+        resp.raise_for_status()
+        data = resp.json()
+        models = data.get("data", [])
+        models.sort(key=lambda m: m.get("id", ""))
+        return [{"id": m["id"], "name": m.get("name", m["id"])} for m in models[:200]]
+
+    else:
+        raise ValueError(f"Unknown provider: {provider}")
+
+
+# ── HTML Pages ────────────────────────────────────────────────────────────
+
+# Shared CSS used by both pages
+SHARED_CSS = """
 *,*::before,*::after{box-sizing:border-box;margin:0;padding:0}
-:root{--bg:#0f172a;--card:#1e293b;--card-hover:#253349;--text:#e2e8f0;--text-dim:#94a3b8;--accent:#f7931a;--green:#22c55e;--red:#ef4444;--yellow:#eab308;--border:#334155;--mono:'JetBrains Mono','Fira Code','Courier New',monospace}
+:root{--bg:#0f172a;--card:#1e293b;--card-hover:#253349;--text:#e2e8f0;--text-dim:#94a3b8;--accent:#f7931a;--green:#22c55e;--red:#ef4444;--yellow:#eab308;--border:#334155;--mono:'JetBrains Mono','Fira Code','Courier New',monospace;--cyan:#22d3ee}
 body{font-family:'Inter',sans-serif;background:var(--bg);color:var(--text);min-height:100vh}
 .container{max-width:1400px;margin:0 auto;padding:16px}
 h1{font-size:1.5rem;font-weight:700;display:flex;align-items:center;gap:10px}
 h1 .btc{color:var(--accent);font-size:1.8rem}
 h2{font-size:1rem;font-weight:600;color:var(--text-dim);margin-bottom:12px;text-transform:uppercase;letter-spacing:.05em;font-size:.8rem}
-
 .header{display:flex;justify-content:space-between;align-items:center;padding:16px 0;border-bottom:1px solid var(--border);margin-bottom:16px;flex-wrap:wrap;gap:12px}
+.nav{display:flex;gap:4px;align-items:center}
+.nav a{color:var(--text-dim);text-decoration:none;font-size:.85rem;font-weight:600;padding:6px 14px;border-radius:6px;transition:all .15s}
+.nav a:hover{color:var(--text);background:var(--card)}
+.nav a.active{color:var(--cyan);background:var(--card);border:1px solid var(--border)}
 .controls{display:flex;gap:8px;align-items:center}
 .btn{padding:8px 18px;border:none;border-radius:6px;font-family:inherit;font-weight:600;font-size:.85rem;cursor:pointer;transition:all .15s}
 .btn-start{background:var(--green);color:#000}.btn-start:hover{background:#16a34a}
 .btn-stop{background:var(--red);color:#fff}.btn-stop:hover{background:#dc2626}
 .btn-secondary{background:var(--border);color:var(--text)}.btn-secondary:hover{background:var(--card-hover)}
+.btn-accent{background:var(--accent);color:#000}.btn-accent:hover{background:#e8850f}
+.btn-cyan{background:var(--cyan);color:#000}.btn-cyan:hover{background:#06b6d4}
 .btn:disabled{opacity:.4;cursor:not-allowed}
-
 .status-badge{display:inline-flex;align-items:center;gap:6px;padding:4px 12px;border-radius:20px;font-size:.8rem;font-weight:600}
 .status-idle{background:#1e3a5f;color:#60a5fa}
 .status-running{background:#1a3a2a;color:var(--green)}
@@ -143,17 +330,57 @@ h2{font-size:1rem;font-weight:600;color:var(--text-dim);margin-bottom:12px;text-
 .status-error{background:#3a1a1a;color:var(--red)}
 .pulse{width:8px;height:8px;border-radius:50%;background:currentColor;animation:pulse 1.5s infinite}
 @keyframes pulse{0%,100%{opacity:1}50%{opacity:.3}}
-
 .best-score{text-align:right}
 .best-score .label{font-size:.7rem;text-transform:uppercase;letter-spacing:.1em;color:var(--text-dim)}
 .best-score .value{font-size:2.2rem;font-weight:700;color:var(--accent);font-family:var(--mono)}
 .best-score .unit{font-size:1rem;color:var(--text-dim)}
+.card{background:var(--card);border-radius:10px;padding:16px;border:1px solid var(--border)}
+.footer{text-align:center;color:var(--text-dim);font-size:.75rem;padding:20px 0;margin-top:16px;border-top:1px solid var(--border)}
+.toast{position:fixed;top:20px;right:20px;padding:12px 20px;border-radius:8px;font-size:.85rem;font-weight:600;z-index:9999;opacity:0;transform:translateY(-10px);transition:all .3s;pointer-events:none}
+.toast.show{opacity:1;transform:translateY(0)}
+.toast-success{background:var(--green);color:#000}
+.toast-error{background:var(--red);color:#fff}
+"""
 
+# Shared HTML head
+SHARED_HEAD = """<meta charset="UTF-8">
+<meta name="viewport" content="width=device-width, initial-scale=1.0">
+<link rel="preconnect" href="https://fonts.googleapis.com">
+<link href="https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700&display=swap" rel="stylesheet">"""
+
+# Navigation bar HTML (parameterized via JS to highlight active page)
+NAV_HTML = """<div class="nav">
+  <a href="/" id="nav-dashboard">Dashboard</a>
+  <a href="/settings" id="nav-settings">&#9881; Settings</a>
+</div>"""
+
+# Toast JS helper (shared)
+TOAST_JS = """
+function showToast(msg, type) {
+  let t = document.getElementById('toast');
+  if (!t) {
+    t = document.createElement('div');
+    t.id = 'toast';
+    t.className = 'toast';
+    document.body.appendChild(t);
+  }
+  t.textContent = msg;
+  t.className = 'toast toast-' + type + ' show';
+  setTimeout(() => { t.classList.remove('show'); }, 3500);
+}
+"""
+
+
+DASHBOARD_HTML = """<!DOCTYPE html>
+<html lang="en">
+<head>
+""" + SHARED_HEAD + """
+<title>BTC Accumulation Signal Optimizer</title>
+<script src="https://cdn.jsdelivr.net/npm/chart.js@4.4.4/dist/chart.umd.min.js"></script>
+<style>
+""" + SHARED_CSS + """
 .grid{display:grid;grid-template-columns:1fr 360px;gap:16px}
 @media(max-width:900px){.grid{grid-template-columns:1fr}}
-
-.card{background:var(--card);border-radius:10px;padding:16px;border:1px solid var(--border)}
-
 .table-wrap{overflow-x:auto;max-height:400px;overflow-y:auto}
 table{width:100%;border-collapse:collapse;font-size:.82rem}
 th{position:sticky;top:0;background:var(--card);text-align:left;padding:8px 10px;color:var(--text-dim);font-weight:600;border-bottom:2px solid var(--border);font-size:.75rem;text-transform:uppercase;letter-spacing:.04em}
@@ -161,13 +388,10 @@ td{padding:7px 10px;border-bottom:1px solid var(--border);font-family:var(--mono
 tr.best-row{background:rgba(34,197,94,.1)}
 tr.best-row td:first-child{border-left:3px solid var(--green)}
 tr:hover{background:var(--card-hover)}
-
 .chart-container{position:relative;height:260px}
-
 .llm-panel{max-height:500px;overflow-y:auto}
 .llm-entry{padding:10px;border-bottom:1px solid var(--border);font-size:.82rem;line-height:1.5}
 .llm-entry .iter-label{font-weight:600;color:var(--accent);font-size:.75rem;margin-bottom:4px}
-
 .config-section{margin-top:16px}
 .config-toggle{cursor:pointer;user-select:none;display:flex;align-items:center;gap:6px}
 .config-toggle .arrow{transition:transform .2s;font-size:.7rem}
@@ -176,12 +400,9 @@ tr:hover{background:var(--card-hover)}
 .config-body.open{display:block}
 textarea.config-editor{width:100%;height:300px;background:var(--bg);color:var(--text);border:1px solid var(--border);border-radius:6px;padding:12px;font-family:var(--mono);font-size:.8rem;resize:vertical}
 .config-actions{display:flex;gap:8px;margin-top:8px}
-
 .downloads{display:flex;gap:8px;margin-top:16px;flex-wrap:wrap}
 .downloads a{color:var(--accent);text-decoration:none;font-size:.82rem;padding:6px 12px;border:1px solid var(--accent);border-radius:6px;transition:all .15s}
 .downloads a:hover{background:var(--accent);color:#000}
-
-.footer{text-align:center;color:var(--text-dim);font-size:.75rem;padding:20px 0;margin-top:16px;border-top:1px solid var(--border)}
 </style>
 </head>
 <body>
@@ -190,14 +411,16 @@ textarea.config-editor{width:100%;height:300px;background:var(--bg);color:var(--
 <div class="header">
   <div>
     <h1><span class="btc">&#x20BF;</span> Accumulation Signal Optimizer</h1>
-    <div style="margin-top:8px">
+    <div style="margin-top:8px;display:flex;align-items:center;gap:12px">
       <span id="statusBadge" class="status-badge status-idle"><span class="pulse"></span> Idle</span>
+      """ + NAV_HTML + """
     </div>
   </div>
   <div style="display:flex;align-items:center;gap:20px;flex-wrap:wrap">
     <div class="controls">
       <button id="btnStart" class="btn btn-start" onclick="startOpt()">Start Optimization</button>
       <button id="btnStop" class="btn btn-stop" onclick="stopOpt()" disabled>Stop</button>
+      <button id="btnClear" class="btn btn-secondary" onclick="clearHistory()" title="Clear all iteration history and start fresh">New Run</button>
     </div>
     <div class="best-score">
       <div class="label">Best Cost Improvement</div>
@@ -262,6 +485,8 @@ textarea.config-editor{width:100%;height:300px;background:var(--bg);color:var(--
 </div>
 
 <script>
+document.getElementById('nav-dashboard').classList.add('active');
+""" + TOAST_JS + """
 let chart = null;
 let pollInterval = null;
 
@@ -324,6 +549,7 @@ function updateStatusBadge(status) {
 
   document.getElementById('btnStart').disabled = (state === 'running');
   document.getElementById('btnStop').disabled = (state !== 'running');
+  document.getElementById('btnClear').disabled = (state === 'running');
   document.getElementById('bestScore').innerHTML = (status.best_score || 0).toFixed(1) + '<span class="unit">%</span>';
 }
 
@@ -406,6 +632,18 @@ async function stopOpt() {
   setTimeout(poll, 500);
 }
 
+async function clearHistory() {
+  if (!confirm('Clear ALL iteration history and start a fresh run? This cannot be undone.')) return;
+  try {
+    const r = await fetch('/api/clear', { method: 'POST' });
+    const d = await r.json();
+    if (d.ok) {
+      document.getElementById('llmPanel').innerHTML = '<div style="color:var(--text-dim);font-size:.82rem;padding:10px">History cleared. Ready for fresh run.</div>';
+      setTimeout(poll, 500);
+    } else { alert('Failed: ' + (d.detail || 'unknown')); }
+  } catch(e) { alert('Failed: ' + e); }
+}
+
 function toggleConfig() {
   const toggle = document.getElementById('configToggle');
   const body = document.getElementById('configBody');
@@ -432,8 +670,8 @@ async function updateConfig() {
       body: JSON.stringify({ config })
     });
     const d = await r.json();
-    if (d.ok) alert('Config updated!');
-  } catch(e) { alert('Invalid JSON or error: ' + e); }
+    if (d.ok) showToast('Config updated!', 'success');
+  } catch(e) { showToast('Invalid JSON or error: ' + e, 'error'); }
 }
 
 async function resetConfig() {
@@ -449,12 +687,350 @@ pollInterval = setInterval(poll, 10000);
 </html>"""
 
 
+SETTINGS_HTML = """<!DOCTYPE html>
+<html lang="en">
+<head>
+""" + SHARED_HEAD + """
+<title>Settings — BTC Accumulation Signal Optimizer</title>
+<style>
+""" + SHARED_CSS + """
+.settings-grid{display:grid;grid-template-columns:320px 1fr;gap:16px;margin-top:16px}
+@media(max-width:800px){.settings-grid{grid-template-columns:1fr}}
+.provider-list{display:flex;flex-direction:column;gap:6px}
+.provider-option{display:flex;align-items:center;gap:10px;padding:12px 14px;border-radius:8px;border:1px solid var(--border);cursor:pointer;transition:all .15s;background:var(--card)}
+.provider-option:hover{border-color:var(--text-dim)}
+.provider-option.selected{border-color:var(--cyan);background:#0f2a3a}
+.provider-option input[type=radio]{accent-color:var(--cyan);width:16px;height:16px}
+.provider-option .provider-name{font-weight:600;font-size:.9rem}
+.provider-option .provider-type{font-size:.7rem;color:var(--text-dim);text-transform:uppercase;letter-spacing:.06em}
+.field-group{margin-bottom:16px}
+.field-group label{display:block;font-size:.75rem;font-weight:600;text-transform:uppercase;letter-spacing:.06em;color:var(--text-dim);margin-bottom:6px}
+.field-group input,.field-group select{width:100%;padding:10px 12px;background:var(--bg);color:var(--text);border:1px solid var(--border);border-radius:6px;font-family:var(--mono);font-size:.85rem}
+.field-group input:focus,.field-group select:focus{outline:none;border-color:var(--cyan)}
+.field-group select{appearance:none;background-image:url("data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' width='12' height='12' viewBox='0 0 12 12'%3E%3Cpath fill='%2394a3b8' d='M6 8L1 3h10z'/%3E%3C/svg%3E");background-repeat:no-repeat;background-position:right 12px center;padding-right:32px}
+.model-select-wrap{position:relative}
+.model-spinner{display:none;position:absolute;right:36px;top:50%;transform:translateY(-50%);width:16px;height:16px;border:2px solid var(--border);border-top-color:var(--cyan);border-radius:50%;animation:spin .6s linear infinite}
+.model-spinner.active{display:block}
+@keyframes spin{to{transform:translateY(-50%) rotate(360deg)}}
+.btn-row{display:flex;gap:8px;margin-top:20px;flex-wrap:wrap}
+.current-provider{font-size:.8rem;color:var(--text-dim);margin-top:4px;font-family:var(--mono)}
+</style>
+</head>
+<body>
+<div class="container">
+
+<div class="header">
+  <div>
+    <h1><span class="btc">&#x20BF;</span> Accumulation Signal Optimizer</h1>
+    <div style="margin-top:8px;display:flex;align-items:center;gap:12px">
+      """ + NAV_HTML + """
+    </div>
+  </div>
+</div>
+
+<div class="card">
+  <h2>&#9881; LLM Provider Settings</h2>
+  <p class="current-provider" id="currentProvider"></p>
+
+  <div class="settings-grid">
+    <div>
+      <h2 style="margin-top:8px">Provider</h2>
+      <div class="provider-list" id="providerList">
+        <label class="provider-option" data-provider="ollama">
+          <input type="radio" name="provider" value="ollama">
+          <div><div class="provider-name">Ollama</div><div class="provider-type">Local</div></div>
+        </label>
+        <label class="provider-option" data-provider="lmstudio">
+          <input type="radio" name="provider" value="lmstudio">
+          <div><div class="provider-name">LM Studio</div><div class="provider-type">Local</div></div>
+        </label>
+        <label class="provider-option" data-provider="openai">
+          <input type="radio" name="provider" value="openai">
+          <div><div class="provider-name">OpenAI</div><div class="provider-type">Cloud</div></div>
+        </label>
+        <label class="provider-option" data-provider="anthropic">
+          <input type="radio" name="provider" value="anthropic">
+          <div><div class="provider-name">Anthropic</div><div class="provider-type">Cloud</div></div>
+        </label>
+        <label class="provider-option" data-provider="openrouter">
+          <input type="radio" name="provider" value="openrouter">
+          <div><div class="provider-name">OpenRouter</div><div class="provider-type">Cloud</div></div>
+        </label>
+      </div>
+    </div>
+
+    <div>
+      <h2 style="margin-top:8px">Connection</h2>
+
+      <div class="field-group" id="fieldBaseUrl" style="display:none">
+        <label>Base URL</label>
+        <input type="text" id="inputBaseUrl" placeholder="http://localhost:11434">
+      </div>
+
+      <div class="field-group" id="fieldApiKey" style="display:none">
+        <label>API Key</label>
+        <input type="password" id="inputApiKey" placeholder="sk-...">
+      </div>
+
+      <div class="field-group">
+        <label>Model</label>
+        <div class="model-select-wrap">
+          <select id="selectModel"><option value="">— select provider first —</option></select>
+          <div class="model-spinner" id="modelSpinner"></div>
+        </div>
+      </div>
+
+      <div class="btn-row">
+        <button class="btn btn-cyan" onclick="testConnection()">Test Connection</button>
+        <button class="btn btn-accent" onclick="saveSettings()">Save Settings</button>
+      </div>
+    </div>
+  </div>
+</div>
+
+<div class="footer">BTC Accumulation Signal Optimizer &mdash; VPS &rarr; Windows GPU &rarr; Mac Mini LLM</div>
+</div>
+
+<script>
+document.getElementById('nav-settings').classList.add('active');
+""" + TOAST_JS + """
+
+let settings = null;
+
+const PROVIDER_FIELDS = {
+  ollama:     { baseUrl: true,  apiKey: false, defaultUrl: 'http://100.100.242.21:11434' },
+  lmstudio:   { baseUrl: true,  apiKey: false, defaultUrl: 'http://100.100.242.21:1234' },
+  openai:     { baseUrl: false, apiKey: true },
+  anthropic:  { baseUrl: false, apiKey: true },
+  openrouter: { baseUrl: false, apiKey: true },
+};
+
+function getSelectedProvider() {
+  const r = document.querySelector('input[name=provider]:checked');
+  return r ? r.value : null;
+}
+
+function buildProviders() {
+  // Build providers dict from current UI state, merging with loaded settings
+  const p = settings ? JSON.parse(JSON.stringify(settings.providers)) : {};
+  const prov = getSelectedProvider();
+  if (!prov) return p;
+
+  if (!p[prov]) p[prov] = {};
+  const fields = PROVIDER_FIELDS[prov];
+  if (fields.baseUrl) {
+    p[prov].base_url = document.getElementById('inputBaseUrl').value;
+  }
+  if (fields.apiKey) {
+    const v = document.getElementById('inputApiKey').value;
+    if (v) p[prov].api_key = v;
+  }
+  return p;
+}
+
+function selectProvider(prov) {
+  // Highlight the selected option
+  document.querySelectorAll('.provider-option').forEach(el => {
+    el.classList.toggle('selected', el.dataset.provider === prov);
+  });
+  document.querySelector('input[name=provider][value="' + prov + '"]').checked = true;
+
+  const fields = PROVIDER_FIELDS[prov];
+  const baseUrlField = document.getElementById('fieldBaseUrl');
+  const apiKeyField = document.getElementById('fieldApiKey');
+
+  baseUrlField.style.display = fields.baseUrl ? 'block' : 'none';
+  apiKeyField.style.display = fields.apiKey ? 'block' : 'none';
+
+  // Populate from settings
+  if (settings && settings.providers[prov]) {
+    const cfg = settings.providers[prov];
+    if (fields.baseUrl) {
+      document.getElementById('inputBaseUrl').value = cfg.base_url || fields.defaultUrl || '';
+    }
+    if (fields.apiKey) {
+      document.getElementById('inputApiKey').value = cfg.api_key || '';
+    }
+  } else {
+    if (fields.baseUrl) document.getElementById('inputBaseUrl').value = fields.defaultUrl || '';
+    if (fields.apiKey) document.getElementById('inputApiKey').value = '';
+  }
+
+  // Reset model dropdown
+  document.getElementById('selectModel').innerHTML = '<option value="">— click Test Connection to load models —</option>';
+}
+
+// Provider click handlers
+document.querySelectorAll('.provider-option').forEach(el => {
+  el.addEventListener('click', () => selectProvider(el.dataset.provider));
+});
+
+async function loadSettings() {
+  try {
+    const r = await fetch('/api/settings');
+    settings = await r.json();
+    document.getElementById('currentProvider').textContent =
+      'Current: ' + settings.provider + ' / ' + settings.model;
+    selectProvider(settings.provider);
+  } catch(e) {
+    console.error('Failed to load settings:', e);
+  }
+}
+
+async function fetchModels(providerOverride) {
+  const prov = providerOverride || getSelectedProvider();
+  if (!prov) return;
+
+  const spinner = document.getElementById('modelSpinner');
+  const sel = document.getElementById('selectModel');
+  spinner.classList.add('active');
+  sel.innerHTML = '<option value="">Loading models...</option>';
+
+  try {
+    const r = await fetch('/api/settings/models', {
+      method: 'POST',
+      headers: { 'Content-Type': 'application/json' },
+      body: JSON.stringify({ provider: prov, providers: buildProviders() })
+    });
+    const data = await r.json();
+    if (!data.ok) {
+      sel.innerHTML = '<option value="">Error: ' + (data.error||'unknown') + '</option>';
+      return;
+    }
+    sel.innerHTML = '';
+    if (!data.models.length) {
+      sel.innerHTML = '<option value="">No models found</option>';
+      return;
+    }
+    for (const m of data.models) {
+      const opt = document.createElement('option');
+      opt.value = m.id;
+      opt.textContent = m.name !== m.id ? m.name + ' (' + m.id + ')' : m.id;
+      sel.appendChild(opt);
+    }
+    // Select current model if it's in the list
+    if (settings && settings.model) {
+      sel.value = settings.model;
+    }
+  } catch(e) {
+    sel.innerHTML = '<option value="">Fetch failed: ' + e + '</option>';
+  } finally {
+    spinner.classList.remove('active');
+  }
+}
+
+async function testConnection() {
+  const prov = getSelectedProvider();
+  if (!prov) { showToast('Select a provider first', 'error'); return; }
+
+  showToast('Testing connection...', 'success');
+  try {
+    const r = await fetch('/api/settings/test', {
+      method: 'POST',
+      headers: { 'Content-Type': 'application/json' },
+      body: JSON.stringify({ provider: prov, providers: buildProviders() })
+    });
+    const data = await r.json();
+    if (data.ok) {
+      showToast(data.message, 'success');
+      // Populate model dropdown
+      const sel = document.getElementById('selectModel');
+      sel.innerHTML = '';
+      for (const m of data.models) {
+        const opt = document.createElement('option');
+        opt.value = m.id;
+        opt.textContent = m.name !== m.id ? m.name + ' (' + m.id + ')' : m.id;
+        sel.appendChild(opt);
+      }
+      if (settings && settings.model) sel.value = settings.model;
+    } else {
+      showToast(data.error || 'Connection failed', 'error');
+    }
+  } catch(e) {
+    showToast('Connection failed: ' + e, 'error');
+  }
+}
+
+async function saveSettings() {
+  const prov = getSelectedProvider();
+  if (!prov) { showToast('Select a provider first', 'error'); return; }
+
+  const model = document.getElementById('selectModel').value;
+  if (!model) { showToast('Select a model first (use Test Connection to load models)', 'error'); return; }
+
+  try {
+    const r = await fetch('/api/settings', {
+      method: 'POST',
+      headers: { 'Content-Type': 'application/json' },
+      body: JSON.stringify({
+        provider: prov,
+        model: model,
+        providers: buildProviders()
+      })
+    });
+    const data = await r.json();
+    if (data.ok) {
+      showToast('Settings saved!', 'success');
+      document.getElementById('currentProvider').textContent = 'Current: ' + prov + ' / ' + model;
+      // Reload settings to get masked keys
+      loadSettings();
+    } else {
+      showToast(data.error || 'Save failed', 'error');
+    }
+  } catch(e) {
+    showToast('Save failed: ' + e, 'error');
+  }
+}
+
+loadSettings();
+</script>
+</body>
+</html>"""
+
+
 @app.get("/", response_class=HTMLResponse)
 def dashboard():
     return DASHBOARD_HTML
 
 
+@app.get("/settings", response_class=HTMLResponse)
+def settings_page():
+    return SETTINGS_HTML
+
+
 if __name__ == "__main__":
     import uvicorn
 
     uvicorn.run(app, host="0.0.0.0", port=3088)
+
+
+@app.post("/api/clear")
+def api_clear():
+    """Clear iteration history for a fresh run."""
+    import glob
+    results_dir = os.path.join(BASE_DIR, "results")
+    
+    # Clear iterations log
+    log_path = os.path.join(results_dir, "iterations.jsonl")
+    if os.path.exists(log_path):
+        os.remove(log_path)
+    
+    # Clear individual result files
+    for f in glob.glob(os.path.join(results_dir, "results_iter_*.json")):
+        os.remove(f)
+    
+    # Reset best config to initial
+    initial = os.path.join(BASE_DIR, "config", "initial_config.json")
+    best = os.path.join(BASE_DIR, "config", "best_config.json")
+    current = os.path.join(BASE_DIR, "config", "current_config.json")
+    if os.path.exists(initial):
+        import shutil
+        shutil.copy(initial, best)
+        shutil.copy(initial, current)
+    
+    # Clear in-memory status
+    orchestrator._status["llm_suggestions"] = []
+    orchestrator._status["best_score"] = 0.0
+    orchestrator._status["iteration"] = 0
+    
+    return {"ok": True, "message": "History cleared. Ready for fresh run."}
diff --git a/llm_client/__pycache__/analyzer.cpython-313.pyc b/llm_client/__pycache__/analyzer.cpython-313.pyc
index 8bf7f7a..28a945a 100644
Binary files a/llm_client/__pycache__/analyzer.cpython-313.pyc and b/llm_client/__pycache__/analyzer.cpython-313.pyc differ
diff --git a/llm_client/analyzer.py b/llm_client/analyzer.py
index 6d0c985..2b6e101 100755
--- a/llm_client/analyzer.py
+++ b/llm_client/analyzer.py
@@ -1,15 +1,36 @@
 #!/usr/bin/env python3
 """
-LLM Accumulation Signal Analyzer -- Calls Ollama on Mac Mini to analyze results
+LLM Accumulation Signal Analyzer -- Calls LLM to analyze results
 and suggest config modifications for the next iteration.
+Supports multiple providers: Ollama, LM Studio, OpenAI, Anthropic, OpenRouter.
 """
 
 import json
+import os
 import re
 import requests
 
-OLLAMA_URL = "http://100.100.242.21:11434"
-MODEL = "qwen3.5:27b"
+BASE_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
+LLM_SETTINGS_PATH = os.path.join(BASE_DIR, "config", "llm_settings.json")
+
+# Fallback defaults
+DEFAULT_OLLAMA_URL = "http://100.100.242.21:11434"
+DEFAULT_MODEL = "qwen3.5:27b"
+
+
+def load_llm_settings():
+    """Load LLM settings from config file, with fallback to defaults."""
+    if os.path.exists(LLM_SETTINGS_PATH):
+        with open(LLM_SETTINGS_PATH) as f:
+            return json.load(f)
+    return {
+        "provider": "ollama",
+        "model": DEFAULT_MODEL,
+        "providers": {
+            "ollama": {"base_url": DEFAULT_OLLAMA_URL},
+        },
+    }
+
 
 SYSTEM_PROMPT = """You are a quantitative analyst optimizing a BTC ACCUMULATION SIGNAL model. The goal is NOT day-trading -- it is finding statistically optimal times to BUY BTC for long-term holding.
 
@@ -120,6 +141,119 @@ You MUST respond with ONLY a JSON object (no markdown, no explanation outside th
 The "config" field must contain the COMPLETE config so it can be used directly."""
 
 
+def _call_ollama(settings, messages):
+    """Call Ollama API."""
+    provider_cfg = settings.get("providers", {}).get("ollama", {})
+    base_url = provider_cfg.get("base_url", DEFAULT_OLLAMA_URL)
+    model = settings.get("model", DEFAULT_MODEL)
+
+    payload = {
+        "model": model,
+        "messages": messages,
+        "stream": False,
+        "think": False,
+        "options": {"temperature": 0.7, "num_predict": 4096},
+    }
+    print(f"  Calling LLM ({model} via Ollama at {base_url})...")
+    resp = requests.post(f"{base_url}/api/chat", json=payload, timeout=600)
+    resp.raise_for_status()
+    return resp.json()["message"]["content"]
+
+
+def _call_openai_compatible(settings, messages, provider_name):
+    """Call OpenAI-compatible API (LM Studio, OpenAI, OpenRouter)."""
+    provider_cfg = settings.get("providers", {}).get(provider_name, {})
+    model = settings.get("model", "")
+
+    if provider_name == "lmstudio":
+        base_url = provider_cfg.get("base_url", "http://100.100.242.21:1234")
+        url = f"{base_url}/v1/chat/completions"
+        headers = {"Content-Type": "application/json"}
+    elif provider_name == "openai":
+        url = "https://api.openai.com/v1/chat/completions"
+        headers = {
+            "Content-Type": "application/json",
+            "Authorization": f"Bearer {provider_cfg.get('api_key', '')}",
+        }
+    elif provider_name == "openrouter":
+        url = "https://openrouter.ai/api/v1/chat/completions"
+        headers = {
+            "Content-Type": "application/json",
+            "Authorization": f"Bearer {provider_cfg.get('api_key', '')}",
+        }
+    else:
+        raise ValueError(f"Unknown OpenAI-compatible provider: {provider_name}")
+
+    payload = {
+        "model": model,
+        "messages": messages,
+        "temperature": 0.7,
+        "max_tokens": 4096,
+    }
+    print(f"  Calling LLM ({model} via {provider_name})...")
+    resp = requests.post(url, json=payload, headers=headers, timeout=600)
+    resp.raise_for_status()
+    return resp.json()["choices"][0]["message"]["content"]
+
+
+def _call_anthropic(settings, messages):
+    """Call Anthropic Messages API."""
+    provider_cfg = settings.get("providers", {}).get("anthropic", {})
+    model = settings.get("model", "claude-sonnet-4-20250514")
+    api_key = provider_cfg.get("api_key", "")
+
+    # Anthropic uses system as a top-level param, not in messages
+    system_msg = ""
+    api_messages = []
+    for m in messages:
+        if m["role"] == "system":
+            system_msg = m["content"]
+        else:
+            api_messages.append(m)
+
+    payload = {
+        "model": model,
+        "max_tokens": 4096,
+        "messages": api_messages,
+    }
+    if system_msg:
+        payload["system"] = system_msg
+
+    headers = {
+        "Content-Type": "application/json",
+        "x-api-key": api_key,
+        "anthropic-version": "2023-06-01",
+    }
+    print(f"  Calling LLM ({model} via Anthropic)...")
+    resp = requests.post(
+        "https://api.anthropic.com/v1/messages",
+        json=payload,
+        headers=headers,
+        timeout=600,
+    )
+    resp.raise_for_status()
+    data = resp.json()
+    # Extract text from content blocks
+    return "".join(
+        block["text"] for block in data.get("content", []) if block.get("type") == "text"
+    )
+
+
+def call_llm(messages):
+    """Route LLM call to the configured provider."""
+    settings = load_llm_settings()
+    provider = settings.get("provider", "ollama")
+
+    if provider == "ollama":
+        return _call_ollama(settings, messages)
+    elif provider in ("lmstudio", "openai", "openrouter"):
+        return _call_openai_compatible(settings, messages, provider)
+    elif provider == "anthropic":
+        return _call_anthropic(settings, messages)
+    else:
+        raise ValueError(f"Unknown LLM provider: {provider}")
+
+
 def analyze_and_suggest(current_config, results, iteration_history=None):
     """
     Send current results to LLM and get suggested config modifications.
@@ -161,24 +295,12 @@ def analyze_and_suggest(current_config, results, iteration_history=None):
 {history_text}
 Analyze these results and suggest 1-3 specific modifications to the config. Return ONLY valid JSON."""
 
-    payload = {
-        "model": MODEL,
-        "messages": [
-            {"role": "system", "content": SYSTEM_PROMPT},
-            {"role": "user", "content": user_prompt},
-        ],
-        "stream": False,
-        "think": False,
-        "options": {
-            "temperature": 0.7,
-            "num_predict": 4096,
-        },
-    }
+    messages = [
+        {"role": "system", "content": SYSTEM_PROMPT},
+        {"role": "user", "content": user_prompt},
+    ]
 
-    print(f"  Calling LLM ({MODEL} on Mac Mini)...")
-    resp = requests.post(f"{OLLAMA_URL}/api/chat", json=payload, timeout=600)
-    resp.raise_for_status()
-    content = resp.json()["message"]["content"]
+    content = call_llm(messages)
 
     # Strip thinking tags if present
     content = re.sub(r"<think>.*?</think>", "", content, flags=re.DOTALL).strip()
@@ -196,7 +318,7 @@ Analyze these results and suggest 1-3 specific modifications to the config. Retu
                 elif content[i] == "}":
                     depth -= 1
                     if depth == 0:
-                        parsed = json.loads(content[brace_start:i + 1])
+                        parsed = json.loads(content[brace_start : i + 1])
                         break
             else:
                 raise ValueError("Could not find complete JSON in LLM response")
@@ -207,7 +329,14 @@ Analyze these results and suggest 1-3 specific modifications to the config. Retu
     changes = parsed.get("changes", [])
     new_config = parsed.get("config", current_config)
 
-    required_keys = ["model_type", "features", "target", "hyperparameters", "strategy", "training"]
+    required_keys = [
+        "model_type",
+        "features",
+        "target",
+        "hyperparameters",
+        "strategy",
+        "training",
+    ]
     for key in required_keys:
         if key not in new_config:
             new_config[key] = current_config[key]
@@ -218,6 +347,7 @@ Analyze these results and suggest 1-3 specific modifications to the config. Retu
 
 if __name__ == "__main__":
     import sys
+
     config_path = sys.argv[1] if len(sys.argv) > 1 else "config/initial_config.json"
     with open(config_path) as f:
         config = json.load(f)
@@ -234,8 +364,18 @@ if __name__ == "__main__":
         "avg_score_at_actual_bottoms": 68.5,
         "avg_score_at_actual_tops": 35.2,
         "per_window_cost_improvement": [7.1, 9.3, 8.8, 10.2, 7.0],
-        "score_distribution": {"0-20": 80, "20-40": 150, "40-60": 200, "60-80": 130, "80-100": 40},
-        "feature_importances": {"dist_from_ath_pct": 0.18, "RSI_14": 0.12, "price_percentile_365": 0.10},
+        "score_distribution": {
+            "0-20": 80,
+            "20-40": 150,
+            "40-60": 200,
+            "60-80": 130,
+            "80-100": 40,
+        },
+        "feature_importances": {
+            "dist_from_ath_pct": 0.18,
+            "RSI_14": 0.12,
+            "price_percentile_365": 0.10,
+        },
     }
 
     new_config, reasoning = analyze_and_suggest(config, dummy_results)
diff --git a/orchestrator.py b/orchestrator.py
index daec0f2..4234a64 100755
--- a/orchestrator.py
+++ b/orchestrator.py
@@ -461,9 +461,22 @@ def run_optimization_loop(callback=None, config_override=None):
                         "iteration": iteration,
                         "reasoning": reasoning,
                     })
+                # Also persist LLM suggestion to iteration log
+                iter_data["llm_reasoning"] = reasoning
+                iter_data["llm_applied"] = True
                 config = new_config
-            except Exception:
-                import random
+            except Exception as e:
+                import random, traceback
+                err_msg = f"LLM call failed: {type(e).__name__}: {e}"
+                print(f"  WARNING: {err_msg}")
+                traceback.print_exc()
+                with _status_lock:
+                    _status["llm_suggestions"].append({
+                        "iteration": iteration,
+                        "reasoning": f"ERROR: {err_msg} — using random perturbation",
+                    })
+                iter_data["llm_reasoning"] = err_msg
+                iter_data["llm_applied"] = False
                 hp = config.get("hyperparameters", {})
                 hp["learning_rate"] = hp.get("learning_rate", 0.01) * random.uniform(0.8, 1.2)
                 hp["max_depth"] = max(3, min(10, hp.get("max_depth", 5) + random.choice([-1, 0, 1])))