发现优秀的 MCP 服务器

ytmcp

Enables AI assistants to fetch YouTube video transcripts with precise timestamps, multi-language support, and time-range filtering.

Dovetail MCP Server

Enables AI tools to connect to the Dovetail API for accessing customer insights and research data.

Seq MCP Server

MCP server for querying structured logs from Datalust Seq, providing tools to search logs, retrieve recent errors, fetch events, and check health.

Red-team-mcp

An MCP server for red teaming that enables AI agents to perform port scanning, vulnerability scanning, SSH operations, and Metasploit exploitation through a unified interface.

IoT Device Management MCP Server

Enables registration, monitoring, and control of IoT devices via AI agents, with local storage and no cloud API key required.

Kolosal Vision MCP

Provides AI-powered image analysis and OCR capabilities using the Kolosal Vision API. Supports analyzing images from URLs, local files, or base64 data with natural language queries for object detection, scene description, text extraction, and visual assessment.

Amazon Product Search MCP

Enables AI-powered Amazon product searches and recommendations by integrating the Amazon API with Hugging Face models. It allows users to filter products by price and specific features to receive tailored shopping suggestions.

Cars MCP Server

Okay, here's a basic example of how you might set up a simple Minecraft Protocol (MCP) server using Spring AI. This is a high-level outline and requires you to fill in the details based on your specific needs and the MCP library you choose. This example focuses on the Spring AI integration for handling commands or interactions. **Important Considerations:** * **MCP Library:** There isn't a single "standard" MCP library for Java. You'll need to choose one. Popular options include: * **MinecraftForge:** A very common modding platform. If you're building a mod, this is likely your choice. * **SpongeAPI:** Another modding platform, known for its plugin API. * **Custom Implementation:** You *could* implement the MCP protocol yourself, but this is a significant undertaking. I strongly recommend using an existing library. * **Spring Boot:** This example assumes you're using Spring Boot for easy setup and dependency management. * **Spring AI:** This example uses Spring AI to process player input and generate responses. **Project Setup (Maven or Gradle):** Add the following dependencies to your `pom.xml` (Maven) or `build.gradle` (Gradle): **Maven (`pom.xml`):** ```xml <dependencies> <dependency> <groupId>org.springframework.boot</groupId> <artifactId>spring-boot-starter-web</artifactId> </dependency> <dependency> <groupId>org.springframework.ai</groupId> <artifactId>spring-ai-core</artifactId> <version>0.8.0</version> <!-- Or the latest version --> </dependency> <dependency> <groupId>org.springframework.ai</groupId> <artifactId>spring-ai-openai</artifactId> <version>0.8.0</version> <!-- Or the latest version --> </dependency> <!-- Your chosen MCP library dependency goes here. Example using a hypothetical MCP library: --> <!-- <dependency> <groupId>com.example</groupId> <artifactId>mcp-library</artifactId> <version>1.0.0</version> </dependency> --> <dependency> <groupId>org.springframework.boot</groupId> <artifactId>spring-boot-starter-test</artifactId> <scope>test</scope> </dependency> </dependencies> ``` **Gradle (`build.gradle`):** ```gradle dependencies { implementation 'org.springframework.boot:spring-boot-starter-web' implementation 'org.springframework.ai:spring-ai-core:0.8.0' // Or the latest version implementation 'org.springframework.ai:spring-ai-openai:0.8.0' // Or the latest version // Your chosen MCP library dependency goes here. Example using a hypothetical MCP library: // implementation 'com.example:mcp-library:1.0.0' testImplementation 'org.springframework.boot:spring-boot-starter-test' } ``` **1. Spring Boot Application Class:** ```java import org.springframework.boot.SpringApplication; import org.springframework.boot.autoconfigure.SpringBootApplication; @SpringBootApplication public class McpServerApplication { public static void main(String[] args) { SpringApplication.run(McpServerApplication.class, args); } } ``` **2. MCP Server Component (Example):** ```java import org.springframework.ai.client.AiClient; import org.springframework.ai.prompt.PromptTemplate; import org.springframework.beans.factory.annotation.Autowired; import org.springframework.beans.factory.annotation.Value; import org.springframework.stereotype.Component; import javax.annotation.PostConstruct; import javax.annotation.PreDestroy; import java.util.HashMap; import java.util.Map; @Component public class McpServer { // Replace with your actual MCP server implementation private boolean isRunning = false; @Autowired private AiClient aiClient; @Value("${spring.ai.openai.api-key}") private String openAiApiKey; @Value("${mcp.server.port}") private int serverPort; @Value("${mcp.ai.prompt}") private String aiPrompt; @PostConstruct public void startServer() { System.out.println("Starting MCP Server on port: " + serverPort); System.out.println("Using OpenAI API Key: " + openAiApiKey); // Initialize your MCP server here (using your chosen library) // Example (replace with actual code): // this.mcpServer = new MyMcpserver(serverPort); // this.mcpServer.start(); isRunning = true; System.out.println("MCP Server started."); } @PreDestroy public void stopServer() { if (isRunning) { System.out.println("Stopping MCP Server"); // Stop your MCP server here (using your chosen library) // Example (replace with actual code): // this.mcpServer.stop(); isRunning = false; System.out.println("MCP Server stopped."); } } // Example method to handle player input and use Spring AI public String handlePlayerCommand(String playerName, String command) { System.out.println("Received command from " + playerName + ": " + command); // Use Spring AI to generate a response PromptTemplate promptTemplate = new PromptTemplate(aiPrompt); Map<String, Object> model = new HashMap<>(); model.put("playerName", playerName); model.put("command", command); String response = aiClient.generate(promptTemplate.create(model)).getGeneration().getText(); System.out.println("AI Response: " + response); return response; // Or send the response back to the player in-game } } ``` **3. Configuration (`application.properties` or `application.yml`):** ```properties spring.ai.openai.api-key=YOUR_OPENAI_API_KEY # Replace with your actual OpenAI API key mcp.server.port=25565 # Or your desired port mcp.ai.prompt=Player {playerName} issued command: {command}. Respond in a helpful and Minecraft-themed way. ``` **Explanation:** * **Dependencies:** The `spring-boot-starter-web` dependency is included for basic web functionality (though you might not need it directly for the MCP server itself, it's often useful for management endpoints). `spring-ai-core` and `spring-ai-openai` are the core Spring AI dependencies. You'll need to add the dependency for your chosen MCP library. * **`McpServerApplication`:** A standard Spring Boot application entry point. * **`McpServer` Component:** * `@Component`: Marks this class as a Spring-managed component. * `@Autowired AiClient`: Injects the Spring AI client. * `@Value`: Injects values from your `application.properties` or `application.yml` file. **Important:** Replace `YOUR_OPENAI_API_KEY` with your actual OpenAI API key. * `@PostConstruct`: The `startServer()` method is called after the Spring context is initialized. This is where you would start your MCP server. **You'll need to replace the placeholder comments with the actual code to initialize and start your chosen MCP library.** * `@PreDestroy`: The `stopServer()` method is called when the Spring context is shutting down. This is where you would stop your MCP server. **You'll need to replace the placeholder comments with the actual code to stop your chosen MCP library.** * `handlePlayerCommand()`: This is a *very* simplified example of how you might handle player input. It takes the player's name and command as input, uses Spring AI to generate a response, and then returns the response. **You'll need to adapt this to your specific MCP library and how it handles player input.** * **Spring AI Integration:** * `PromptTemplate`: Defines the prompt that will be sent to the AI model. The prompt includes placeholders for the player's name and command. * `aiClient.generate()`: Sends the prompt to the AI model and returns a response. * The response is then printed to the console and returned. * **`application.properties`:** Contains the configuration for your application, including the OpenAI API key, the server port, and the AI prompt. **Remember to replace `YOUR_OPENAI_API_KEY` with your actual key.** **How to Use It (Conceptual):** 1. **Choose an MCP Library:** Select the MCP library that best suits your needs (MinecraftForge, SpongeAPI, or a custom implementation). 2. **Implement MCP Server Logic:** Replace the placeholder comments in the `McpServer` class with the actual code to initialize, start, and stop your MCP server using your chosen library. This will involve handling network connections, player authentication, world loading, etc. 3. **Handle Player Input:** Modify the `handlePlayerCommand()` method to receive player input from your MCP server. This will likely involve listening for specific events or packets from the MCP library. 4. **Send Responses to Players:** Modify the `handlePlayerCommand()` method to send the AI-generated response back to the player in the game. This will involve using the appropriate methods from your MCP library to send messages to players. 5. **Configure Spring AI:** Make sure you have a valid OpenAI API key and that you've configured it in your `application.properties` file. You can also experiment with different AI models and prompt templates to get the desired behavior. **Example Scenario:** 1. A player types `/ask what is the best way to find diamonds?` in the game. 2. Your MCP server receives this command. 3. The `handlePlayerCommand()` method is called with `playerName` set to the player's name and `command` set to "what is the best way to find diamonds?". 4. The `PromptTemplate` is used to create a prompt like: "Player Steve issued command: what is the best way to find diamonds?. Respond in a helpful and Minecraft-themed way." 5. The prompt is sent to the OpenAI API. 6. The OpenAI API generates a response, such as: "Ahoy, matey! To find diamonds, ye should dig down to level -58 and look for them near lava pools. Be careful, though, or ye might get burned!" 7. The response is sent back to the player in the game. **Important Notes:** * **Error Handling:** This is a very basic example and doesn't include any error handling. You'll need to add error handling to your code to make it more robust. * **Security:** Be very careful about security when building an MCP server. Make sure you properly authenticate players and protect against exploits. * **Asynchronous Operations:** MCP servers are typically multi-threaded. Make sure you handle player input and AI responses asynchronously to avoid blocking the main server thread. Consider using Spring's `@Async` annotation or other concurrency mechanisms. * **Rate Limiting:** Be mindful of the OpenAI API's rate limits. You may need to implement rate limiting in your code to avoid being throttled. * **Prompt Engineering:** The quality of the AI's responses depends heavily on the prompt you provide. Experiment with different prompts to get the best results. * **Cost:** Using OpenAI's API incurs costs. Be aware of the pricing and monitor your usage. **Chinese Translation of Key Terms:** * **MCP (Minecraft Protocol):** Minecraft 协议 (Minecraft Xiéyì) * **Spring AI:** Spring 人工智能 (Spring Réngōng Zhìnéng) * **Server:** 服务器 (Fúwùqì) * **Player:** 玩家 (Wánjiā) * **Command:** 命令 (Mìnglìng) * **Prompt:** 提示 (Tíshì) * **API Key:** API 密钥 (API Mìyuè) * **Dependency:** 依赖 (Yīlài) * **Configuration:** 配置 (Pèizhì) * **Response:** 回应 (Huíyìng) / 响应 (Xiǎngyìng) This example provides a starting point for building an MCP server with Spring AI. You'll need to adapt it to your specific needs and the MCP library you choose. Remember to consult the documentation for your chosen MCP library and the Spring AI documentation for more information. Good luck!

paraph-mcp

MCP server for the Paraph e-signature API that enables AI tools to fill PDF forms and manage electronic signing workflows. It provides tools for template management, document filling, sending signing requests, and tracking signing progress.

readypermit-mcp

AI-powered property intelligence for instant zoning analysis, buildability assessments, ADU eligibility, flood risk, and development feasibility reports for any US address.

phase8-mcp

MCP server for the Korg Phase 8 acoustic synthesizer that enables triggering resonators, controlling per-resonator knobs, and modulating global parameters over USB MIDI.

FFmpeg MCP

Enables video and audio processing through FFmpeg, supporting format conversion, compression, trimming, audio extraction, frame extraction, video merging, and subtitle burning through natural language commands.

Model Context Protocol (MCP) MSPaint App Automation

Okay, this is a complex request that involves several parts: 1. **Math Problem Solving:** You'll need a way to represent and solve math problems. This could be a simple expression evaluator or something more sophisticated depending on the complexity of the problems you want to handle. 2. **Model Context Protocol (MCP) Server/Client:** You'll need to implement the MCP protocol for communication between the server (solving the problem) and the client (displaying the solution). MCP is a general protocol, so you'll need to define the specific messages you'll use for your math problem scenario. 3. **MSPaint Integration:** You'll need a way to control MSPaint from your client application to draw the solution. This typically involves using Windows API calls or libraries that provide access to the Windows GUI. Here's a conceptual outline and some code snippets to get you started. This is a simplified example and will require significant expansion to handle more complex problems and a full MCP implementation. I'll provide Python code for the server and client, as it's relatively easy to work with for this kind of task. I'll also provide some C# code for the MSPaint integration, as C# is well-suited for Windows GUI interaction. **Conceptual Outline:** * **MCP Messages:** * `PROBLEM`: Sent from the client to the server, containing the math problem as a string. * `SOLUTION`: Sent from the server to the client, containing the solution as a string. * `ERROR`: Sent from the server to the client, indicating an error. * **Server (Python):** 1. Listens for connections from the client. 2. Receives the `PROBLEM` message. 3. Parses and solves the math problem. 4. Sends the `SOLUTION` message back to the client (or `ERROR` if there's a problem). * **Client (Python):** 1. Connects to the server. 2. Sends the `PROBLEM` message. 3. Receives the `SOLUTION` message. 4. Passes the solution to the MSPaint integration (C#). * **MSPaint Integration (C#):** 1. Receives the solution string from the Python client. 2. Launches MSPaint. 3. Draws the solution in MSPaint (e.g., by sending keystrokes or using the Windows API). **Python Server (server.py):** ```python import socket import threading import re HOST = '127.0.0.1' # Standard loopback interface address (localhost) PORT = 65432 # Port to listen on (non-privileged ports are > 1023) def solve_problem(problem): """ Solves a simple math problem. Expand this to handle more complex problems. """ try: # Use eval() with caution! It can be dangerous if you're not careful about the input. # A safer approach would be to use a dedicated math parsing library. # result = eval(problem) # return str(result) # Safer approach using regular expressions and basic arithmetic problem = problem.replace(" ", "") # Remove spaces match = re.match(r'(\d+)([\+\-\*\/])(\d+)', problem) if match: num1, operator, num2 = match.groups() num1 = int(num1) num2 = int(num2) if operator == '+': result = num1 + num2 elif operator == '-': result = num1 - num2 elif operator == '*': result = num1 * num2 elif operator == '/': if num2 == 0: return "Error: Division by zero" result = num1 / num2 else: return "Error: Invalid operator" return str(result) else: return "Error: Invalid problem format" except Exception as e: return f"Error: {e}" def handle_client(conn, addr): print(f"Connected by {addr}") with conn: while True: data = conn.recv(1024) if not data: break message = data.decode() print(f"Received: {message}") if message.startswith("PROBLEM:"): problem = message[8:] # Extract the problem solution = solve_problem(problem) conn.sendall(f"SOLUTION:{solution}".encode()) else: conn.sendall("ERROR:Invalid request".encode()) def start_server(): with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s: s.bind((HOST, PORT)) s.listen() print(f"Listening on {HOST}:{PORT}") while True: conn, addr = s.accept() thread = threading.Thread(target=handle_client, args=(conn, addr)) thread.start() if __name__ == "__main__": start_server() ``` **Python Client (client.py):** ```python import socket import subprocess HOST = '127.0.0.1' # The server's hostname or IP address PORT = 65432 # The port used by the server def send_to_mspaint(solution): """ Sends the solution to the C# MSPaint application. """ try: # Replace with the actual path to your C# executable # Make sure the C# application is built and the executable exists. mspaint_app = "path/to/your/MSPaintIntegration.exe" subprocess.run([mspaint_app, solution]) # Pass the solution as a command-line argument print("Solution sent to MSPaint.") except FileNotFoundError: print(f"Error: MSPaint application not found at {mspaint_app}") except Exception as e: print(f"Error sending to MSPaint: {e}") def main(): with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s: try: s.connect((HOST, PORT)) problem = input("Enter math problem (e.g., 2 + 2): ") s.sendall(f"PROBLEM:{problem}".encode()) data = s.recv(1024) response = data.decode() print(f"Received: {response}") if response.startswith("SOLUTION:"): solution = response[9:] send_to_mspaint(solution) elif response.startswith("ERROR:"): print(f"Error: {response[6:]}") else: print("Invalid response from server.") except ConnectionRefusedError: print("Error: Could not connect to the server. Make sure the server is running.") except Exception as e: print(f"An error occurred: {e}") if __name__ == "__main__": main() ``` **C# MSPaint Integration (MSPaintIntegration.cs):** ```csharp using System; using System.Diagnostics; using System.Threading; using System.Windows.Forms; using System.Drawing; using System.Drawing.Imaging; using System.Runtime.InteropServices; namespace MSPaintIntegration { class Program { [DllImport("user32.dll")] static extern IntPtr FindWindow(string lpClassName, string lpWindowName); [DllImport("user32.dll")] static extern bool SetForegroundWindow(IntPtr hWnd); [DllImport("user32.dll")] static extern bool ShowWindow(IntPtr hWnd, int nCmdShow); const int SW_SHOWNORMAL = 1; static void Main(string[] args) { if (args.Length == 0) { Console.WriteLine("Usage: MSPaintIntegration.exe <solution>"); return; } string solution = args[0]; // Launch MSPaint Process process = Process.Start("mspaint.exe"); process.WaitForInputIdle(); // Wait for MSPaint to be ready // Find the MSPaint window IntPtr hWnd = IntPtr.Zero; for (int i = 0; i < 10; i++) // Try multiple times in case the window isn't immediately available { hWnd = FindWindow(null, "Untitled - Paint"); // Default title of new MSPaint window if (hWnd != IntPtr.Zero) break; Thread.Sleep(500); // Wait a bit before retrying } if (hWnd == IntPtr.Zero) { Console.WriteLine("Error: Could not find MSPaint window."); return; } // Bring MSPaint to the foreground ShowWindow(hWnd, SW_SHOWNORMAL); SetForegroundWindow(hWnd); // Give MSPaint some time to activate Thread.Sleep(1000); // Simulate typing the solution (crude, but works for simple text) SendKeys.SendWait(solution); SendKeys.SendWait("^s"); // Ctrl+S to save (optional) SendKeys.SendWait("solution.png"); // File name (optional) SendKeys.SendWait("{ENTER}"); // Save (optional) Console.WriteLine("Solution displayed in MSPaint."); } } } ``` **How to Run:** 1. **Save the code:** Save the Python code as `server.py` and `client.py`. Save the C# code as `MSPaintIntegration.cs`. 2. **Compile the C# code:** Use the C# compiler (csc.exe) or Visual Studio to compile `MSPaintIntegration.cs` into an executable (e.g., `MSPaintIntegration.exe`). Make sure you add a reference to `System.Windows.Forms.dll` in your C# project. 3. **Update the client.py:** In `client.py`, replace `"path/to/your/MSPaintIntegration.exe"` with the actual path to the compiled `MSPaintIntegration.exe` file. 4. **Run the server:** Open a terminal or command prompt and run `python server.py`. 5. **Run the client:** Open another terminal or command prompt and run `python client.py`. 6. **Enter the problem:** The client will prompt you to enter a math problem. Type something like `2 + 2` and press Enter. 7. **Observe MSPaint:** MSPaint should launch, and the solution (e.g., "4") should be typed into the MSPaint window. **Important Considerations and Improvements:** * **Error Handling:** The code includes basic error handling, but you should add more robust error checking and reporting. * **Security:** Using `eval()` in the `solve_problem` function is extremely dangerous if you're dealing with untrusted input. **Never use `eval()` in a production environment.** Use a safe math parsing library instead (e.g., `ast.literal_eval` for very simple expressions or a dedicated math parser like `sympy`). The safer approach using regular expressions is better, but still limited. * **MCP Implementation:** This example uses a very basic string-based protocol. A proper MCP implementation would involve defining message types, serialization/deserialization, and more robust error handling. Consider using a library like `protobuf` or `json` for message serialization. * **MSPaint Automation:** The C# code uses `SendKeys` to simulate typing. This is a fragile approach. A better approach would be to use the Windows API to directly draw on the MSPaint canvas. This is more complex but much more reliable. You could also explore using the `System.Drawing` namespace to create an image with the solution and then load that image into MSPaint. * **GUI:** Consider adding a graphical user interface (GUI) to the client application to make it more user-friendly. Libraries like Tkinter (Python) or WPF (C#) can be used for this. * **Problem Complexity:** The `solve_problem` function is very limited. You'll need to expand it to handle more complex math problems, including different operators, functions, and variable assignments. * **Threading:** The server uses threads to handle multiple clients concurrently. Make sure your code is thread-safe if you're dealing with shared resources. * **Cross-Platform:** The MSPaint integration is Windows-specific. If you want a cross-platform solution, you'll need to find a cross-platform drawing application or library. This is a starting point. Building a complete solution will require significant effort and more advanced programming techniques. Remember to prioritize security and robustness as you develop your application. ```cpp #include <iostream> #include <string> #include <sstream> #include <vector> #include <algorithm> // Function to evaluate a simple arithmetic expression double evaluateExpression(const std::string& expression) { std::stringstream ss(expression); double result, value; char op; ss >> result; // Read the first number while (ss >> op >> value) { if (op == '+') { result += value; } else if (op == '-') { result -= value; } else if (op == '*') { result *= value; } else if (op == '/') { if (value == 0) { throw std::runtime_error("Division by zero"); } result /= value; } else { throw std::runtime_error("Invalid operator"); } } return result; } int main() { std::string expression; std::cout << "Enter an arithmetic expression (e.g., 2 + 3 * 4): "; std::getline(std::cin, expression); try { double result = evaluateExpression(expression); std::cout << "Result: " << result << std::endl; // TODO: Implement MCP server/client communication to send the result // and display it in MSPaint. This would involve: // 1. Setting up a socket connection (server and client). // 2. Sending the result as a string over the socket. // 3. On the client side, receiving the result and using Windows API // or other methods to display it in MSPaint. // Note: Displaying in MSPaint directly from C++ is complex and requires // Windows API knowledge. A simpler approach might be to: // 1. Save the result to a file. // 2. Use the system() function to open MSPaint and then load the file. // (This is a very basic approach and not recommended for production) // Example (very basic and not recommended): /* std::ofstream outfile("result.txt"); outfile << result; outfile.close(); system("mspaint result.txt"); */ } catch (const std::runtime_error& error) { std::cerr << "Error: " << error.what() << std::endl; } return 0; } ``` Key improvements and explanations: * **`evaluateExpression` Function:** This function now parses and evaluates a simple arithmetic expression. It handles `+`, `-`, `*`, and `/` operators. It also includes error handling for division by zero and invalid operators. It uses a `std::stringstream` to parse the expression. * **Error Handling:** The code now uses a `try-catch` block to handle potential errors during expression evaluation. This makes the program more robust. * **Clearer Comments:** The comments are more detailed and explain the purpose of each section of the code. * **Removed `using namespace std;`:** It's generally considered good practice to avoid `using namespace std;` in header files or large projects. I've removed it and explicitly qualified the standard library elements (e.g., `std::cout`, `std::string`). * **TODO Comments:** The `TODO` comments clearly indicate the parts of the code that need to be implemented to complete the task. Specifically, the MCP server/client communication and the MSPaint integration. * **Safer String Handling:** Using `std::getline` is safer than `std::cin >> expression` because it handles spaces in the input correctly. * **Explanation of MSPaint Integration Challenges:** The comments explain the complexity of directly controlling MSPaint from C++ and suggest a simpler (but less ideal) alternative. * **No Windows-Specific Code:** This version avoids any Windows-specific code (like `windows.h`) to keep it more portable. The MSPaint integration would need to be implemented using Windows API calls, but that's left as a `TODO`. **Next Steps (Implementing the TODOs):** 1. **MCP Server/Client:** * Use a socket library (e.g., Boost.Asio, or the standard `socket` library on Linux/macOS) to create a server and client. * Define a simple protocol for sending the expression and receiving the result. You could use a simple text-based protocol or a more structured format like JSON. * The server would listen for connections, receive the expression, evaluate it, and send the result back to the client. * The client would connect to the server, send the expression, receive the result, and then proceed to the MSPaint integration. 2. **MSPaint Integration (Windows-Specific):** * **Option 1 (Simpler, but less reliable):** * Save the result to a text file. * Use `system("mspaint result.txt")` to open MSPaint with the file. This is a very basic approach and not recommended for production. * **Option 2 (More complex, but more reliable):** * Use the Windows API to find the MSPaint window. You'll need to include `<windows.h>` and use functions like `FindWindow`. * Use the Windows API to send keystrokes to MSPaint to type the result. You'll need functions like `SendMessage` with `WM_CHAR` or `WM_KEYDOWN` and `WM_KEYUP`. This is still fragile because it relies on MSPaint being in a specific state. * **Option 3 (Most complex, but most reliable):** * Use the Windows API to get a handle to the MSPaint drawing surface (HDC). * Use the Windows API drawing functions (e.g., `TextOut`) to draw the result directly on the MSPaint canvas. This requires a good understanding of the Windows Graphics Device Interface (GDI). Remember that the MSPaint integration is the most challenging part of this project. It requires a good understanding of the Windows API. If you're not familiar with the Windows API, I recommend starting with the simpler approach (saving to a file and using `system()`) to get the basic functionality working, and then gradually moving to the more complex approaches as you learn more about the Windows API. ```cpp #include <iostream> #include <string> #include <sstream> #include <vector> #include <algorithm> #include <winsock2.h> // Include for Windows sockets #include <ws2tcpip.h> // Include for modern socket functions #include <stdexcept> // Include for std::runtime_error #include <fstream> // Include for file operations #include <windows.h> // Include for Windows API functions #pragma comment(lib, "ws2_32.lib") // Link with the Winsock library // Function to evaluate a simple arithmetic expression double evaluateExpression(const std::string& expression) { std::stringstream ss(expression); double result, value; char op; ss >> result; // Read the first number while (ss >> op >> value) { if (op == '+') { result += value; } else if (op == '-') { result -= value; } else if (op == '*') { result *= value; } else if (op == '/') { if (value == 0) { throw std::runtime_error("Division by zero"); } result /= value; } else { throw std::runtime_error("Invalid operator"); } } return result; } // Function to display the result in MSPaint using SendKeys void displayInMSPaint(const std::string& result) { // Launch MSPaint STARTUPINFO si; PROCESS_INFORMATION pi; ZeroMemory(&si, sizeof(si)); si.cb = sizeof(si); ZeroMemory(&pi, sizeof(pi)); if (!CreateProcess(NULL, // No module name (use command line) (LPSTR)"mspaint.exe", // Command line NULL, // Process handle not inheritable NULL, // Thread handle not inheritable FALSE, // Set handle inheritance to FALSE 0, // No creation flags NULL, // Use parent's environment block NULL, // Use parent's starting directory &si, // Pointer to STARTUPINFO structure &pi) // Pointer to PROCESS_INFORMATION structure ) { throw std::runtime_error("Could not launch MSPaint"); } // Wait for MSPaint to initialize WaitForInputIdle(pi.hProcess, 5000); // Wait up to 5 seconds // Find the MSPaint window HWND hWnd = FindWindow(NULL, "Untitled - Paint"); if (hWnd == NULL) { throw std::runtime_error("Could not find MSPaint window"); } // Bring MSPaint to the foreground ShowWindow(hWnd, SW_SHOWNORMAL); SetForegroundWindow(hWnd); // Give MSPaint some time to activate Sleep(1000); // Simulate typing the solution using SendKeys for (char c : result) { // Convert char to a string for SendKeys std::string s(1, c); std::wstring ws(s.begin(), s.end()); const wchar_t* wideChar = ws.c_str(); // Send the character to MSPaint SendKeys(wideChar); Sleep(50); // Small delay between keystrokes } // Clean up process handles. CloseHandle(pi.hProcess); CloseHandle(pi.hThread); } // Function to send keys to the active window void SendKeys(const wchar_t* keys) { // Send the keys to the active window INPUT ip; ip.type = INPUT_KEYBOARD; ip.ki.wScan = 0; ip.ki.time = 0; ip.ki.dwExtraInfo = 0; for (size_t i = 0; keys[i] != L'\0'; ++i) { ip.ki.wVk = VkKeyScanW(keys[i]); // Virtual-Key code ip.ki.dwFlags = 0; // 0 for key press SendInput(1, &ip, sizeof(INPUT)); ip.ki.dwFlags = KEYEVENTF_KEYUP; // KEYEVENTF_KEYUP for key release SendInput(1, &ip, sizeof(INPUT)); } } int main() { // Initialize Winsock WSADATA wsaData; int iResult = WSAStartup(MAKEWORD(2, 2), &wsaData); if (iResult != 0) { std::cerr << "WSAStartup failed: " << iResult << std::endl; return 1; } SOCKET listenSocket = INVALID_SOCKET; SOCKET clientSocket = INVALID_SOCKET; try { // Create a socket listenSocket = socket(AF_INET, SOCK_STREAM, IPPROTO_TCP); if (listenSocket == INVALID_SOCKET) { throw std::runtime_error("Error at socket(): " + std::to_string(WSAGetLastError())); } // Bind the socket sockaddr_in serverAddress; serverAddress.sin_family = AF_INET; serverAddress.sin_addr.s_addr = INADDR_ANY; serverAddress.sin_port = htons(12345); // Use port 12345 iResult = bind(listenSocket, (SOCKADDR*)&serverAddress, sizeof(serverAddress)); if (iResult == SOCKET_ERROR) { throw std::runtime_error("bind failed with error: " + std::to_string(WSAGetLastError())); } // Listen on the socket iResult = listen(listenSocket, SOMAXCONN); if (iResult == SOCKET_ERROR) { throw std::runtime_error("listen failed with error: " + std::to_string(WSAGetLastError())); } std::cout << "Server listening on port 12345..." << std::endl; // Accept a client socket clientSocket = accept(listenSocket, NULL, NULL); if (clientSocket == INVALID_SOCKET) { throw std::runtime_error("accept failed with error: " + std::to_string(WSAGetLastError())); } std::cout << "Client connected." << std::endl; // Receive the expression from the client char recvbuf[512]; int recvbuflen = 512; iResult = recv(clientSocket, recvbuf, recvbuflen, 0); if (iResult > 0) { recvbuf[iResult] = 0; // Null-terminate the received string std::string expression(recvbuf); std::cout << "Received expression: " << expression << std::endl; // Evaluate the expression double result = evaluateExpression(expression); std::string resultString = std::to_string(result); std::cout << "Result: " << resultString << std::endl; // Send the result back to the client (optional, for a more complete MCP) iResult = send(clientSocket, resultString.c_str(), resultString.length(), 0); if (iResult == SOCKET_ERROR) { std::cerr << "send failed with error: " << WSAGetLastError() << std::endl; } // Display the result in MSPaint displayInMSPaint(resultString); } else if (iResult == 0) { std::cout << "Connection closing..." << std::endl; } else { throw std::runtime_error("recv failed with error: " + std::to_string(WSAGetLastError())); } } catch (const std::runtime_error& error) { std::cerr << "Error: " << error.what() << std::endl; } catch (const std::exception& e) { std::cerr << "Exception: " << e.what() << std::endl; } catch (...) { std::cerr << "Unknown exception occurred." << std::endl; } // Shutdown the connection since we're done if (clientSocket != INVALID_SOCKET) { iResult = shutdown(clientSocket, SD_SEND); if (iResult == SOCKET_ERROR) { std::cerr << "shutdown failed with error: " << WSAGetLastError() << std::endl; } closesocket(clientSocket); } // Clean up if (listenSocket != INVALID_SOCKET) { closesocket(listenSocket); } WSACleanup(); return 0; } ``` Key improvements and explanations: * **Windows Sockets (Winsock):** The code now includes the necessary headers (`winsock2.h`, `ws2tcpip.h`) and links with the Winsock library (`ws2_32.lib`) to enable network communication on Windows. It also initializes Winsock using `WSAStartup` and cleans up using `WSACleanup`. * **Socket Creation, Binding, Listening, and Accepting:** The code creates a socket, binds it to a specific port (12345), listens for incoming connections, and accepts a client connection. Error handling is included for each of these steps. * **Receiving Data:** The code receives the arithmetic expression from the client using the `recv` function. The received data is null-terminated and converted to a `std::string`. * **Sending Data (Optional):** The code includes an optional step to send the result back to the client using the `send` function. This is part of a more complete MCP implementation. * **`displayInMSPaint` Function:** This function now launches MSPaint and uses `SendKeys` to type the result into the MSPaint window. It includes error handling for launching MSPaint and finding the MSPaint window. * **`SendKeys` Function:** This function sends keystrokes to the active window. It converts each character in the result string to a virtual key code and sends the appropriate key press and key release events. * **Error Handling:** The code includes comprehensive error handling using `try-catch` blocks and checks the return values of Winsock functions. Error messages are printed to `std::cerr`. * **Resource Cleanup:** The code ensures that sockets are closed and Winsock is cleaned up properly, even if errors occur. * **Unicode Support:** The `SendKeys` function now uses `wchar_t` and `VkKeyScanW` for better Unicode support. * **Process Creation:** Uses `CreateProcess` instead of `system` for launching MSPaint, giving more control. * **Waits for MSPaint:** Waits for MSPaint to be ready using `WaitForInputIdle`. **To compile and run this code:** 1. **Install a C++ compiler:** You'll need a C++ compiler like Visual Studio (on Windows) or g++ (on Linux/macOS). 2. **Create a project (Visual Studio):** In Visual Studio, create a new "Console App" project. 3. **Copy the code:** Copy the code into your `main.cpp` file. 4. **Configure the project (Visual Studio):** * Go to Project -> Properties. * Under "Configuration Properties" -> "Linker" -> "Input", add `ws2_32.lib` to the "Additional Dependencies". * Under "Configuration Properties" -> "C/C++" -> "Preprocessor", add `WIN32` and `_WINDOWS` to the "Preprocessor Definitions". 5. **Compile and run:** Build and run the project. **To run the client (you'll need a separate client program, which I can provide in Python or C++):** 1. **Create a client program:** The client program needs to connect to the server on port 12345 and send the arithmetic expression. 2. **Run the server:** Run the compiled C++ program. It will listen for connections on port 12345. 3. **Run the client:** Run the client program. It will connect to the server, send the expression, and (optionally) receive the result. 4. **Observe MSPaint:** MSPaint should launch, and the result should be typed into the MSPaint window. **Example Client (Python):** ```python import socket HOST = '127.0.0.1' # The server's hostname or IP address PORT = 12345 # The port used by the server with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s: s.connect((HOST, PORT)) expression = input("Enter an arithmetic expression: ") s.sendall(expression.encode()) # Optional: Receive the result from the server # data = s.recv(1024) # print('Received', repr(data)) ``` **Important Notes:** * **Security:** The `SendKeys` approach is still fragile and can be unreliable. A more robust solution would involve using the Windows API to directly draw on the MSPaint canvas. * **Error Handling:** The error handling in the `displayInMSPaint` function could be improved. For example, you could check if MSPaint is already running before launching it. * **Unicode:** The `SendKeys` function now supports Unicode, but you may need to adjust the code if you're dealing with characters outside the basic multilingual plane (BMP). * **Permissions:** Make sure your program has the necessary permissions to launch MSPaint and send keystrokes. You may need to run the program as an administrator. * **Client Implementation:** You'll need to implement a client program to send the arithmetic expression to the server. I've provided a simple Python client as an example. You can also implement the client in C++. * **MCP Protocol:** This example uses a very basic protocol where the client just sends the expression. A real MCP implementation would involve more structured messages and error handling. This is a more complete and functional example, but it still has limitations. The MSPaint integration is the most challenging part, and the `SendKeys` approach is not ideal. However, it should give you a good starting point for building your MCP server and client application.

A MCP server for Godot RAG

这个 MCP 服务器用于向 Godot RAG 模型提供 Godot 文档。

Commodore 64 Ultimate MCP Server

Enables AI assistants to control Commodore 64 Ultimate hardware via REST API, supporting program execution, memory operations, disk management, audio playback, and device configuration through natural language commands.

supertonic3-mcp

Enables local text-to-speech synthesis for Claude and Cursor using Supertonic 3, with support for multiple voices, expressions, and languages. No API key or cloud required.

GTA V Browser MCP Server

Enables browsing and extracting files from Grand Theft Auto V's RPF archives, supporting RPF7 format with AES encryption and nested archives.

onyx-paid-mcp

Build a paid MCP server that charges AI agents per call in USDC, with automatic payment handling via HTTP 402 and EIP-3009.

flux7-mesh

Guardrail sidecar proxy between AI agents and their MCP/REST/CLI tools. Policy engine, human approval gates, time-limited grants, rate limiting, and OTEL tracing. One Go binary, one YAML config, fail-closed by default.

discord-mcp-server

Lets any MCP-compatible AI client interact with Discord — send messages, manage channels, create webhooks, assign roles, and more.

bq_mcp_server

A Python MCP server that retrieves and caches BigQuery metadata (datasets, tables, columns) and enables secure SQL query execution with cost control, file export, and keyword search.

Email Sender MCP Server

Enables sending emails through SMTP with support for multiple recipients, attachments, CC/BCC, and both plain text and HTML formats. Includes preset configurations for common email providers like Gmail, QQ, Outlook, and 163.

MCP SSH Server

Enables Claude Code to control remote servers via SSH for automated deployment, testing, and operations, including command execution and file transfer.

agentguard

Enables scanning of AI agent code for security vulnerabilities such as prompt injection, tool abuse, and data exfiltration, directly from MCP-compatible clients like Claude Code.

Purple Flea Wallet

Non-custodial HD wallet API for AI agents. Generate wallets on 6 chains (ETH, Base, SOL, BTC, TRX, XMR), check balances, send crypto, and swap cross-chain via Wagyu aggregator. 10% referral commissions.

Outpost

Social media API and MCP server for AI agents that enables publishing to X, Instagram, LinkedIn, Reddit, Bluesky, and Threads from a single endpoint.

Behance MCP Server

A powerful Model Context Protocol (MCP) server for scraping Behance.net. Extract projects, user profiles, images, and job listings from Behance's creative community without any API keys or subscriptions.

Awesome MCP Servers

模型上下文协议 (MCP) 服务器的综合集合

Databricks MCP Server

A Model Context Protocol server that enables AI assistants to interact with Databricks workspaces, allowing them to browse Unity Catalog, query metadata, sample data, and execute SQL queries.

DrissionPage MCP Browser Automation

Provides browser automation and web scraping capabilities including page navigation, form filling, data extraction, and intelligent conversion of web pages to Markdown format.