diff --git a/CHANGELOG.md b/CHANGELOG.md
index 03efa2d..827ef4b 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -30,7 +30,7 @@ Peekaboo MCP is now production-ready! This release marks the culmination of exte
 - Robust permission handling for Screen Recording and Accessibility
 
 ### Requirements
-- macOS 15.0 or later (Sequoia)
+- macOS 14.0 or later (Sonoma)
 - Node.js 18 or later
 - Screen Recording permission (for capture features)
 - Accessibility permission (optional, for foreground window detection)
diff --git a/README.md b/README.md
index 0fe7e47..0122bf4 100644
--- a/README.md
+++ b/README.md
@@ -4,7 +4,7 @@
 
 [![npm version](https://badge.fury.io/js/%40steipete%2Fpeekaboo-mcp.svg)](https://www.npmjs.com/package/@steipete/peekaboo-mcp)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
-[![macOS](https://img.shields.io/badge/macOS-15.0%2B-blue.svg)](https://www.apple.com/macos/)
+[![macOS](https://img.shields.io/badge/macOS-14.0%2B-blue.svg)](https://www.apple.com/macos/)
 [![Node.js](https://img.shields.io/badge/node-%3E%3D20.0.0-brightgreen.svg)](https://nodejs.org/)
 
 Peekaboo is a macOS-only MCP server that enables AI agents to capture screenshots of applications, windows, or the entire system, with optional visual question answering through local or remote AI models.
@@ -33,7 +33,7 @@ Read more about the design philosophy and implementation details in the [blog po
 
 ### Requirements
 
-- **macOS 15.0+** (Sequoia or later)
+- **macOS 14.0+** (Sonoma or later)
 - **Node.js 20.0+**
 - **Screen Recording Permission** (you'll be prompted on first use)
 
diff --git a/peekaboo-cli/Package.swift b/peekaboo-cli/Package.swift
index dc17558..9959bbc 100644
--- a/peekaboo-cli/Package.swift
+++ b/peekaboo-cli/Package.swift
@@ -4,7 +4,7 @@ import PackageDescription
 let package = Package(
     name: "peekaboo",
     platforms: [
-        .macOS(.v15)
+        .macOS(.v14)
     ],
     products: [
         .executable(
diff --git a/peekaboo-cli/Sources/peekaboo/main.swift b/peekaboo-cli/Sources/peekaboo/main.swift
index d0d60e5..88bb133 100644
--- a/peekaboo-cli/Sources/peekaboo/main.swift
+++ b/peekaboo-cli/Sources/peekaboo/main.swift
@@ -2,7 +2,7 @@ import ArgumentParser
 import Foundation
 
 @main
-@available(macOS 10.15, *)
+@available(macOS 14.0, *)
 struct PeekabooCommand: AsyncParsableCommand {
     static let configuration = CommandConfiguration(
         commandName: "peekaboo",
diff --git a/src/utils/ai-providers.ts b/src/utils/ai-providers.ts
index 4f12060..3afb1d6 100644
--- a/src/utils/ai-providers.ts
+++ b/src/utils/ai-providers.ts
@@ -106,6 +106,9 @@ async function analyzeWithOllama(
 
   logger.debug({ model, baseUrl }, "Analyzing image with Ollama");
 
+  // Default to describing the image if no question is provided
+  const prompt = question.trim() || "Please describe what you see in this image.";
+
   const response = await fetch(`${baseUrl}/api/generate`, {
     method: "POST",
     headers: {
@@ -113,7 +116,7 @@ async function analyzeWithOllama(
     },
     body: JSON.stringify({
       model,
-      prompt: question,
+      prompt,
       images: [imageBase64],
       stream: false,
     }),
@@ -147,13 +150,16 @@ async function analyzeWithOpenAI(
 
   const openai = new OpenAI({ apiKey });
 
+  // Default to describing the image if no question is provided
+  const prompt = question.trim() || "Please describe what you see in this image.";
+
   const response = await openai.chat.completions.create({
     model: model || "gpt-4o",
     messages: [
       {
         role: "user",
         content: [
-          { type: "text", text: question },
+          { type: "text", text: prompt },
           {
             type: "image_url",
             image_url: {
diff --git a/tests/unit/tools/analyze-edge-cases.test.ts b/tests/unit/tools/analyze-edge-cases.test.ts
index 81db6c4..d7a267a 100644
--- a/tests/unit/tools/analyze-edge-cases.test.ts
+++ b/tests/unit/tools/analyze-edge-cases.test.ts
@@ -37,7 +37,7 @@ describe("Analyze Tool - Edge Cases", () => {
   });
 
   describe("Empty question handling", () => {
-    it("should handle empty string question", async () => {
+    it("should handle empty string question with default prompt", async () => {
       mockParseAIProviders.mockReturnValue([
         { provider: "ollama", model: "llava:latest" }
       ]);
@@ -49,8 +49,8 @@ describe("Analyze Tool - Edge Cases", () => {
       
       mockReadImageAsBase64.mockResolvedValue("base64imagedata");
       
-      // Mock Ollama returning "No response from Ollama" for empty question
-      mockAnalyzeImageWithProvider.mockResolvedValue("No response from Ollama");
+      // Mock Ollama returning a description of the image
+      mockAnalyzeImageWithProvider.mockResolvedValue("This image shows a desktop window with various UI elements.");
 
       const result = await analyzeToolHandler(
         { 
@@ -69,11 +69,11 @@ describe("Analyze Tool - Edge Cases", () => {
       );
 
       expect(result.content[0].type).toBe("text");
-      expect(result.content[0].text).toBe("No response from Ollama");
+      expect(result.content[0].text).toBe("This image shows a desktop window with various UI elements.");
       expect(result.model_used).toBe("ollama/llava:latest");
     });
 
-    it("should handle whitespace-only question", async () => {
+    it("should handle whitespace-only question with default prompt", async () => {
       mockParseAIProviders.mockReturnValue([
         { provider: "ollama", model: "llava:latest" }
       ]);
@@ -84,7 +84,7 @@ describe("Analyze Tool - Edge Cases", () => {
       });
       
       mockReadImageAsBase64.mockResolvedValue("base64imagedata");
-      mockAnalyzeImageWithProvider.mockResolvedValue("No response from Ollama");
+      mockAnalyzeImageWithProvider.mockResolvedValue("This image shows a screenshot of an application.");
 
       const result = await analyzeToolHandler(
         { 
@@ -102,10 +102,10 @@ describe("Analyze Tool - Edge Cases", () => {
         mockLogger,
       );
 
-      expect(result.content[0].text).toBe("No response from Ollama");
+      expect(result.content[0].text).toBe("This image shows a screenshot of an application.");
     });
 
-    it("should handle question with only newlines", async () => {
+    it("should handle question with only newlines with default prompt", async () => {
       mockParseAIProviders.mockReturnValue([
         { provider: "ollama", model: "llava:latest" }
       ]);
@@ -116,7 +116,7 @@ describe("Analyze Tool - Edge Cases", () => {
       });
       
       mockReadImageAsBase64.mockResolvedValue("base64imagedata");
-      mockAnalyzeImageWithProvider.mockResolvedValue("No response from Ollama");
+      mockAnalyzeImageWithProvider.mockResolvedValue("This image displays a user interface with multiple sections.");
 
       const result = await analyzeToolHandler(
         { 
@@ -134,7 +134,7 @@ describe("Analyze Tool - Edge Cases", () => {
         mockLogger,
       );
 
-      expect(result.content[0].text).toBe("No response from Ollama");
+      expect(result.content[0].text).toBe("This image displays a user interface with multiple sections.");
     });
   });
 
diff --git a/tests/unit/utils/ai-providers.test.ts b/tests/unit/utils/ai-providers.test.ts
index 353b5c0..2bbeb39 100644
--- a/tests/unit/utils/ai-providers.test.ts
+++ b/tests/unit/utils/ai-providers.test.ts
@@ -335,6 +335,51 @@ describe("AI Providers Utility", () => {
       expect(result).toBe("No response from Ollama");
     });
 
+    it("should use default prompt for empty question with Ollama", async () => {
+      (global.fetch as vi.Mock).mockResolvedValueOnce({
+        ok: true,
+        json: async () => ({ response: "This image shows a window with text content." }),
+      });
+      const result = await analyzeImageWithProvider(
+        { provider: "ollama", model: "llava" },
+        "path/img.png",
+        imageBase64,
+        "", // Empty question
+        mockLogger,
+      );
+      expect(result).toBe("This image shows a window with text content.");
+      const fetchCall = (global.fetch as vi.Mock).mock.calls[0];
+      const body = JSON.parse(fetchCall[1].body);
+      expect(body.prompt).toBe("Please describe what you see in this image.");
+    });
+
+    it("should use default prompt for whitespace-only question with OpenAI", async () => {
+      process.env.OPENAI_API_KEY = "test-key";
+      mockChatCompletionsCreate.mockResolvedValueOnce({
+        choices: [{ message: { content: "This image displays a user interface." } }],
+      });
+
+      const result = await analyzeImageWithProvider(
+        { provider: "openai", model: "gpt-4o" },
+        "path/img.png",
+        imageBase64,
+        "   ", // Whitespace-only question
+        mockLogger,
+      );
+      expect(result).toBe("This image displays a user interface.");
+      expect(mockChatCompletionsCreate).toHaveBeenCalledWith(
+        expect.objectContaining({
+          messages: expect.arrayContaining([
+            expect.objectContaining({
+              content: expect.arrayContaining([
+                { type: "text", text: "Please describe what you see in this image." },
+              ]),
+            }),
+          ]),
+        }),
+      );
+    });
+
     it("should throw error for anthropic provider (not implemented)", async () => {
       await expect(
         analyzeImageWithProvider(