Feat: Improve Model Tier Recommendation for Apple Silicon

backend/app/utils/hardware_detect.py

@@ -1,5 +1,8 @@
+from __future__ import annotations
+
 import shutil
 import subprocess
+import platform
 
 import onnxruntime as ort
 import psutil
@@ -35,11 +38,60 @@ def detect_physical_gpu() -> list[str]:
     return gpu_names
 
 
+def detect_apple_silicon() -> str | None:
+    if platform.system() != "Darwin":
+        return None
+    if platform.machine() != "arm64":
+        return None
+
+    sysctl = shutil.which("sysctl")
+    if sysctl:
+        try:
+            result = subprocess.run(
+                [sysctl, "-n", "machdep.cpu.brand_string"],
+                capture_output=True,
+                text=True,
+                check=True,
+                timeout=5,
+            )
+            chip_name = result.stdout.strip()
+            if chip_name:
+                return chip_name
+        except (OSError, subprocess.SubprocessError):
+            pass
+
+    proc = platform.processor()
+    if proc:
+        return proc
+
+    return "Apple Silicon"
+
+
+def detect_apple_silicon_tier() -> str | None:
+    chip = detect_apple_silicon()
+    if chip is None:
+        return None
+
+    chip_lower = chip.lower()
+
+    # Base M1 and M2 chips (without Pro/Max/Ultra modifiers) are "small"
+    if any(base in chip_lower for base in ["m1", "m2"]):
+        if not any(mod in chip_lower for mod in ["pro", "max", "ultra"]):
+            return "small"
+
+    # Default to medium for all other Apple Silicon (Pro/Max/Ultra, M3+, and future chips)
+    return "medium"
+
+
 def detect_hardware_tier() -> str:
     """
     Detect system hardware to recommend the best YOLO/FaceNet model tier.
     Returns: 'nano', 'small', or 'medium'
     """
+    apple_tier = detect_apple_silicon_tier()
+    if apple_tier is not None:
+        return apple_tier
+
     # Check RAM in GB
     ram_gb = psutil.virtual_memory().total / (1024**3)
 
@@ -66,6 +118,7 @@ def get_hardware_info() -> dict:
         "ram_gb": round(psutil.virtual_memory().total / (1024**3), 2),
         "gpu_detected": bool(gpu_names),
         "gpu_names": gpu_names,
+        "apple_silicon": detect_apple_silicon(),
         "available_providers": ort.get_available_providers(),
         "recommended_tier": detect_hardware_tier(),
     }

frontend/src/components/OnboardingSteps/AIModelSetupStep.tsx

@@ -262,9 +262,11 @@ export const AIModelSetupStep: React.FC<AIModelSetupStepProps> = ({
                   </span>
                   <span className="bg-muted/60 text-muted-foreground inline-flex items-center gap-1 rounded-full border px-3 py-1 text-xs font-medium">
                     <Cpu className="h-3 w-3" />
-                    {hardwareSpecs.gpu_names?.length
-                      ? hardwareSpecs.gpu_names.join(', ')
-                      : 'No dedicated GPU'}
+                    {hardwareSpecs.apple_silicon
+                      ? hardwareSpecs.apple_silicon
+                      : hardwareSpecs.gpu_names?.length
+                        ? hardwareSpecs.gpu_names.join(', ')
+                        : 'No dedicated GPU'}
                   </span>
                 </div>
               )}

frontend/src/types/models.ts

@@ -24,6 +24,7 @@ export interface HardwareInfo {
   ram_gb: number;
   gpu_detected: boolean;
   gpu_names: string[];
+  apple_silicon: string | null;
   available_providers: string[];
   recommended_tier: ModelTier;
 }