Cache Provider-Modelle in models_cache.json statt bei jedem Start von APIs zu laden

- LoadModelsAsync() prüft zuerst data/models_cache.json
- Bei Cache-Hit: kein API-Call, sofortiger Start
- RefreshModelsFromApis() lädt von APIs + speichert Cache
- POST /api/models/refresh Endpoint für manuellen Refresh
- dataDir an OpenAiService-Konstruktor übergeben

Services/OpenAiService.cs

@@ -1,5 +1,6 @@
 using System;
 using System.Collections.Generic;
+using System.IO;
 using System.Net.Http;
 using System.Net.Http.Headers;
 using System.Text;
@@ -151,12 +152,14 @@ public class OpenAiService
     private List<ProviderModel> _nvidiaModels      = new List<ProviderModel>();
     private List<ProviderModel> _pollinationsModels = new List<ProviderModel>();
     private List<ProviderModel> _geminiModels       = new List<ProviderModel>();
+    private readonly string _cacheFile;
 
-    public OpenAiService(IHttpClientFactory httpFactory, DatabaseService db, ToolService toolService, Microsoft.Extensions.Configuration.IConfiguration configuration)
+    public OpenAiService(IHttpClientFactory httpFactory, DatabaseService db, ToolService toolService, Microsoft.Extensions.Configuration.IConfiguration configuration, string dataDir)
     {
         _httpFactory = httpFactory;
         _db = db;
         _toolService = toolService;
+        _cacheFile = Path.Combine(dataDir, "models_cache.json");
         Providers = new Dictionary<string, (string BaseUrl, string ApiKey)>
         {
             ["nvidia"]      = ("https://g4f.space/api/nvidia/v1",                                    ""),
@@ -166,19 +169,65 @@ public class OpenAiService
     }
 
     /// <summary>
-    /// Lädt verfügbare Modelle aller Provider beim Server-Start und cached sie.
+    /// Lädt verfügbare Modelle – aus lokalem Cache (models_cache.json) oder von den APIs.
     /// </summary>
-    public async Task LoadModelsAsync()
+    public async Task LoadModelsAsync(bool forceRefresh = false)
+    {
+        if (!forceRefresh && File.Exists(_cacheFile))
+        {
+            try
+            {
+                var json = await File.ReadAllTextAsync(_cacheFile);
+                var cache = JsonSerializer.Deserialize<Dictionary<string, List<ProviderModel>>>(json);
+                if (cache != null)
+                {
+                    _nvidiaModels       = cache.ContainsKey("nvidia")       ? cache["nvidia"]       : new List<ProviderModel>();
+                    _pollinationsModels = cache.ContainsKey("pollinations") ? cache["pollinations"] : new List<ProviderModel>();
+                    _geminiModels       = cache.ContainsKey("gemini")       ? cache["gemini"]       : new List<ProviderModel>();
+                    Console.WriteLine($"[OpenAiService] Modelle aus Cache geladen: {_nvidiaModels.Count} Nvidia, {_pollinationsModels.Count} Pollinations, {_geminiModels.Count} Gemini");
+                    return;
+                }
+            }
+            catch (Exception ex)
+            {
+                Console.WriteLine($"[OpenAiService] Cache-Datei fehlerhaft, lade von APIs: {ex.Message}");
+            }
+        }
+
+        await RefreshModelsFromApis();
+    }
+
+    /// <summary>
+    /// Lädt Modelle von allen Provider-APIs und speichert sie im Cache.
+    /// </summary>
+    public async Task RefreshModelsFromApis()
     {
         _nvidiaModels       = await FetchModels("https://g4f.space/api/nvidia/models", "Nvidia");
         _pollinationsModels = await FetchModels("https://text.pollinations.ai/models", "Pollinations");
 
-        // Gemini: dynamisch von der API laden (Bearer-Token erforderlich)
         var geminiApiKey = Providers["gemini"].ApiKey;
         if (!string.IsNullOrEmpty(geminiApiKey))
             _geminiModels = await FetchModels("https://generativelanguage.googleapis.com/v1beta/openai/models", "Gemini", geminiApiKey);
         else
             Console.WriteLine("[OpenAiService] GeminiApiKey nicht konfiguriert – Gemini-Modelle nicht geladen.");
+
+        // Cache speichern
+        try
+        {
+            var cache = new Dictionary<string, List<ProviderModel>>
+            {
+                ["nvidia"]       = _nvidiaModels,
+                ["pollinations"] = _pollinationsModels,
+                ["gemini"]       = _geminiModels
+            };
+            var json = JsonSerializer.Serialize(cache, new JsonSerializerOptions { WriteIndented = true });
+            await File.WriteAllTextAsync(_cacheFile, json);
+            Console.WriteLine($"[OpenAiService] Modelle-Cache gespeichert: {_cacheFile}");
+        }
+        catch (Exception ex)
+        {
+            Console.WriteLine($"[OpenAiService] Fehler beim Speichern des Modelle-Cache: {ex.Message}");
+        }
     }
 
     private async Task<List<ProviderModel>> FetchModels(string url, string providerName, string apiKey = "")