feat: multi-provider AI support with per-provider model selection

- Add llm-factory.ts: unified LLM provider abstraction (Google/Claude/OpenAI)
- Install @langchain/anthropic and @langchain/openai packages
- resolveAiProvider picks active provider from DB settings (Claude > OpenAI > Google)
- All AI graphs (reminder, new-patient, reschedule, internal-chat) now accept provider+model params
- Add claudeAiModel, openAiModel, googleAiModel columns to ai_settings table
- New PUT /api/ai/provider-model route to save selected model per provider
- UI model dropdowns for Claude (Haiku/Sonnet/Opus), OpenAI (GPT-5.x series), Google (Gemini 2.5/3.x)
- Google AI section also gets model selector alongside existing API key field

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

apps/Backend/src/routes/ai-settings.ts

@@ -2,6 +2,7 @@ import express, { Request, Response } from "express";
 import { storage } from "../storage";
 import { classifyInternalChat } from "../ai/internal-chat-graph";
 import { runInternalChatWorkflow } from "../ai/internal-chat-workflow";
+import { resolveAiProvider } from "../ai/llm-factory";
 
 const router = express.Router();
 
@@ -20,8 +21,11 @@ router.get("/settings", async (req: Request, res: Response): Promise<any> => {
       aiEnabled:        settings.aiEnabled ?? true,
       openAiKey:        settings.openAiKey ?? "",
       openAiEnabled:    settings.openAiEnabled ?? false,
+      openAiModel:      settings.openAiModel ?? "gpt-5.2",
       claudeAiKey:      settings.claudeAiKey ?? "",
       claudeAiEnabled:  settings.claudeAiEnabled ?? false,
+      claudeAiModel:    settings.claudeAiModel ?? "claude-haiku-4-5-20251001",
+      googleAiModel:    settings.googleAiModel ?? "gemini-2.5-flash",
       dentalMgmtKey:    settings.dentalMgmtKey ?? "",
       dentalMgmtEnabled: settings.dentalMgmtEnabled ?? false,
       openPhoneReply:   settings.openPhoneReply ?? false,
@@ -109,6 +113,27 @@ router.put("/provider-enabled", async (req: Request, res: Response): Promise<any
   }
 });
 
+// PUT /api/ai/provider-model
+router.put("/provider-model", async (req: Request, res: Response): Promise<any> => {
+  try {
+    const userId = req.user?.id;
+    if (!userId) return res.status(401).json({ message: "Unauthorized" });
+
+    const { provider, model } = req.body;
+    if (!["claudeAi", "openAi", "googleAi"].includes(provider)) {
+      return res.status(400).json({ message: "Invalid provider" });
+    }
+    if (!model?.trim()) {
+      return res.status(400).json({ message: "model is required" });
+    }
+
+    await storage.setProviderModel(userId, provider, model.trim());
+    return res.status(200).json({ provider, model: model.trim() });
+  } catch (err) {
+    return res.status(500).json({ error: "Failed to save provider model", details: String(err) });
+  }
+});
+
 // GET /api/ai/advanced-settings
 router.get("/advanced-settings", async (req: Request, res: Response): Promise<any> => {
   try {
@@ -236,9 +261,10 @@ router.post("/internal-chat", async (req: Request, res: Response): Promise<any>
     if (!message?.trim()) return res.status(400).json({ message: "message is required" });
 
     const aiSettings = await storage.getAiSettings(userId);
-    if (!aiSettings?.apiKey) {
+    const activeAi   = resolveAiProvider(aiSettings ?? {});
+    if (!activeAi) {
       return res.status(200).json({
-        reply: "AI is not configured. Please add a Google AI API key in Settings.",
+        reply: "AI is not configured. Please add an API key in AI Settings.",
       });
     }
 
@@ -249,9 +275,11 @@ router.post("/internal-chat", async (req: Request, res: Response): Promise<any>
 
     const classification = await classifyInternalChat(
       message.trim(),
-      aiSettings.apiKey,
+      activeAi.key,
       extraSystemPrompt || undefined,
-      Array.isArray(history) ? history : []
+      Array.isArray(history) ? history : [],
+      activeAi.provider,
+      activeAi.model
     );
 
     const response = await runInternalChatWorkflow(classification, userId, storage, customAliases);