diff --git a/package.json b/package.json
index 4d916aa9..bc06c2f8 100644
--- a/package.json
+++ b/package.json
@@ -3,7 +3,7 @@
   "version": "1.0.0",
   "private": true,
   "scripts": {
-    "dev": "next dev",
+    "dev": "node server.js",
     "build": "next build",
     "start": "next start",
     "lint": "next lint",
diff --git a/server.js b/server.js
new file mode 100644
index 00000000..cfa096b1
--- /dev/null
+++ b/server.js
@@ -0,0 +1,30 @@
+// Custom Next.js server with extended timeout for video generation
+// Node.js default server.requestTimeout is 5 minutes (300,000ms)
+// We extend it to 10 minutes for long-running fal.ai video generation
+
+const { createServer } = require('http');
+const next = require('next');
+
+const dev = process.env.NODE_ENV !== 'production';
+const hostname = 'localhost';
+const port = process.env.PORT || 3000;
+
+const app = next({ dev, hostname, port });
+const handle = app.getRequestHandler();
+
+app.prepare().then(() => {
+  const server = createServer(async (req, res) => {
+    // Use WHATWG URL API instead of deprecated url.parse()
+    const parsedUrl = new URL(req.url, `http://${hostname}:${port}`);
+    await handle(req, res, parsedUrl);
+  });
+
+  // Increase timeout to 10 minutes for long-running video generation
+  server.requestTimeout = 600000; // 10 minutes
+  server.headersTimeout = 610000; // Slightly longer than requestTimeout
+
+  server.listen(port, () => {
+    console.log(`> Ready on http://${hostname}:${port}`);
+    console.log(`> Server timeout set to ${server.requestTimeout / 1000 / 60} minutes`);
+  });
+});
diff --git a/src/app/api/generate/route.ts b/src/app/api/generate/route.ts
index fe5c3418..ae8f34b4 100644
--- a/src/app/api/generate/route.ts
+++ b/src/app/api/generate/route.ts
@@ -1,10 +1,23 @@
+/**
+ * Generate API Route
+ * 
+ * TIMEOUT CONFIGURATION:
+ * - maxDuration: Only applies on Vercel, not locally
+ * - AbortSignal.timeout: Controls outgoing fetch to providers
+ * - For local development, server.requestTimeout must be set in server.js (Node.js default is 5 minutes)
+ * 
+ * FAL.AI QUEUE API NOTE:
+ * The generateWithFalQueue function exists but is NOT used because fal.ai's queue API
+ * has file size limitations that are too restrictive for our use case. We use the blocking
+ * fal.run endpoint instead, which requires the server timeout to be extended for video generation.
+ */
 import { NextRequest, NextResponse } from "next/server";
 import { GoogleGenAI } from "@google/genai";
 import { GenerateRequest, GenerateResponse, ModelType, SelectedModel, ProviderType } from "@/types";
 import { GenerationInput, GenerationOutput, ProviderModel } from "@/lib/providers/types";
 import { uploadImageForUrl, shouldUseImageUrl, deleteImages } from "@/lib/images";
 
-export const maxDuration = 600; // 10 minute timeout for video generation
+export const maxDuration = 600; // 10 minute timeout for video generation (Vercel only)
 export const dynamic = 'force-dynamic'; // Ensure this route is always dynamic
 
 // Map model types to Gemini model IDs
@@ -889,7 +902,11 @@ async function generateWithFal(
 /**
  * Generate video using fal.ai Queue API
  * Uses async queue submission + polling to handle long-running video generation
- * that would otherwise timeout with the blocking fal.run endpoint
+ * that would otherwise timeout with the blocking fal.run endpoint.
+ * 
+ * NOTE: This function is NOT currently used because fal.ai's queue API has file size
+ * limitations that are too restrictive. We use the blocking fal.run endpoint instead
+ * with an extended server timeout configured in server.js.
  */
 async function generateWithFalQueue(
   requestId: string,