Discover Awesome MCP Servers

Explore MCP

A demonstration MCP server that exposes basic arithmetic tools (add, subtract, ping) through FastAPI and shows how to integrate them with OpenAI's tool-calling API for LLM orchestration.

origin-MCP

Enables AI assistants to control Origin 2025b for scientific plotting via natural language, with support for data import, 52+ chart types, curve fitting, statistics, and export.

Habit Tracker AI MCP

Habit Tracker AI - MCP server providing AI-powered tools and automation by MEOK AI Labs

gitequity-mcp

MCP server that exposes your GitEquity portfolio to AI clients, enabling balance queries, stock prices, and trade proposals confirmed via GitHub comments.

Datadog Logs MCP Server

Enables searching and retrieving Datadog logs through the Model Context Protocol with customizable queries, time ranges, and result limits.

arduino-mcp-server

Server MCP Arduino yang ditulis dalam Go.

rcw-mcp

An MCP server for the Revised Code of Washington (RCW) that enables retrieving full statutory text by citation, searching for citations by partial reference, and finding bills that affect a given citation.

AgentLadle MCP AKShare

Provides 46 financial data tools for AI assistants covering A-share, HK, US markets, macroeconomics, funds, and derivatives, powered by AKShare.

koios-mcp

An MCP server that provides LLMs with access to 95 tools covering the Koios Cardano blockchain REST API, enabling queries for on-chain data like transactions, addresses, assets, and governance.

HREVN MCP Server

Minimal stdio MCP server that exposes HREVN compliance and audit tools as structured MCP tools, enabling baseline checks, profile validation, and bundle generation via a managed runtime.

CashClaw GHL MCP

An MCP server for GoHighLevel with 82 live-tested tools, enabling CRM operations like contact management, appointments, invoices, and workflows via natural language.

SimuBridge

Enables AI assistants to control MATLAB Simulink models through natural language, providing tools for model creation, block management, wiring, simulation, and more via a local MCP backend.

mcp-dev-tools

Provides AI coding assistants with tools to run git status, recent commits, tests, and linter on a real repository.

MojaWave MCP

Connects any MCP-compatible AI assistant to the MojaWave SMS Gateway, enabling sending SMS, checking credit balances, and managing bulk SMS jobs.

xtapdown-mcp

An MCP server that provides LLM clients with direct access to XTapDown's Twitter creator toolkit, enabling tweet downloads, engagement calculations, content generation, and trend analysis without authentication or rate limits.

hwms-mcp-server

AI-driven module selection and scaffold generation for hybrid web applications. Enables automatic dependency resolution and project structure creation via natural language queries.

MCP Server

Server MCP

mcp-a2a-documentation

Search and retrieve Agent2Agent (A2A) protocol documentation using full-text search and section filtering.

Phpactor MCP Server

Provides project-wide semantic PHP refactoring tools (find references, rename classes/members, move class, class info) for AI assistants by wrapping Phpactor.

supragents-mcp

A read-only research MCP server that provides search and browsing tools for Hacker News, Reddit, and Product Hunt. Works with zero API keys for basic use.

AWS Documentation MCP Server

Enables users to access, search, and get recommendations from AWS documentation through natural language queries. Supports both global AWS documentation and AWS China documentation with tools to fetch pages, search content, and discover related resources.

upnote-mcp

Enables AI assistants to interact with Upnote via its x-callback-url API, allowing creation of notes, notebooks, tag management, and search. It also supports navigation to various Upnote sections and custom filters.

Acopia

MCP server for querying and simulating the dispatch plan of a solar PV + battery system in the Chilean electricity market, using deterministic optimization and optional DRL.

ai.cabrini/market-data

Provides US stock market data for AI agents, including intraday and daily bars, SEC fundamentals, filings, and insider data, with pay-per-query via USDC on Base.

agenticpay

agenticpay lets MCP server developers monetize tools via per-call USDC micropayments on Solana, using the x402 protocol. Each tool declares a price; agents pay via signed Solana transactions; settlement happens on-chain in ~1.5–2 seconds.

mcp-4o-Image-Generator

mcp-4o-Image-Generator

nucleo-mcp

Enables AI assistants to search, retrieve, and save Nucleo icons from the user's local library directly into projects.

Zyxel Switch MCP Server

Enables AI applications to interact with Zyxel switches via CLI for network configuration, monitoring, and management, supporting SSH/Telnet sessions and MCP-compliant tools, resources, and prompts.

User Info MCP Server

An MCP server providing tools for user information management with capabilities for retrieving, searching, and adding user data stored in a JSON file.

Mcp Server

Okay, here's an example of a simple MCP (Minecraft Protocol) server written in Python, designed to be easily understood and potentially adapted for use with a large language model like Claude. This is a very basic example and doesn't implement all the features of a real Minecraft server. It focuses on handling the initial handshake and sending a simple status response. ```python import socket import struct import json def handle_handshake(sock): """Handles the initial handshake from the client.""" # Read the packet length (VarInt) packet_length, bytes_read = read_varint(sock) print(f"Packet length: {packet_length}, Bytes read: {bytes_read}") # Read the packet ID (VarInt) packet_id, bytes_read_id = read_varint(sock) print(f"Packet ID: {packet_id}, Bytes read: {bytes_read_id}") # Read the protocol version (VarInt) protocol_version, bytes_read_version = read_varint(sock) print(f"Protocol Version: {protocol_version}, Bytes read: {bytes_read_version}") # Read the server address (String) server_address_length, bytes_read_address_length = read_varint(sock) server_address = sock.recv(server_address_length).decode('utf-8') print(f"Server Address: {server_address}, Bytes read: {bytes_read_address_length}") # Read the server port (Unsigned Short) server_port = struct.unpack('>H', sock.recv(2))[0] # Big-endian unsigned short print(f"Server Port: {server_port}") # Read the next state (VarInt) next_state, bytes_read_state = read_varint(sock) print(f"Next State: {next_state}, Bytes read: {bytes_read_state}") return next_state def handle_status_request(sock): """Handles the status request from the client and sends a response.""" # Read the packet length (VarInt) - should be 1 for an empty status request packet_length, bytes_read = read_varint(sock) print(f"Status Request Packet Length: {packet_length}, Bytes read: {bytes_read}") # Read the packet ID (VarInt) - should be 0 for a status request packet_id, bytes_read_id = read_varint(sock) print(f"Status Request Packet ID: {packet_id}, Bytes read: {bytes_read_id}") # Construct the status response status = { "version": { "name": "My Awesome Server", "protocol": 763 # Example protocol version (1.17.1) }, "players": { "max": 100, "online": 0, "sample": [] }, "description": { "text": "A server powered by Python and AI!" } } status_json = json.dumps(status) status_bytes = status_json.encode('utf-8') status_length = len(status_bytes) # Create the response packet packet = bytearray() write_varint(packet, status_length) # Length of the JSON string packet.extend(status_bytes) # Prepend the packet length packet_length = len(packet) response = bytearray() write_varint(response, packet_length) response.extend(packet) # Send the response sock.sendall(response) def handle_ping(sock): """Handles the ping request from the client and sends a response.""" # Read the packet length (VarInt) - should be 9 packet_length, bytes_read = read_varint(sock) print(f"Ping Packet Length: {packet_length}, Bytes read: {bytes_read}") # Read the packet ID (VarInt) - should be 1 packet_id, bytes_read_id = read_varint(sock) print(f"Ping Packet ID: {packet_id}, Bytes read: {bytes_read_id}") # Read the payload (Long) payload = struct.unpack('>q', sock.recv(8))[0] # Big-endian long print(f"Ping Payload: {payload}") # Create the response packet (same payload) response = bytearray() write_varint(response, 8) # Packet Length (8 bytes for long) write_varint(response, 0) # Packet ID (0 for pong) response.extend(struct.pack('>q', payload)) # Payload # Prepend the packet length packet_length = len(response) final_response = bytearray() write_varint(final_response, packet_length) final_response.extend(response) # Send the response sock.sendall(final_response) def read_varint(sock): """Reads a VarInt from the socket.""" num_read = 0 result = 0 shift = 0 while True: byte = sock.recv(1)[0] num_read += 1 result |= (byte & 0x7F) << shift shift += 7 if not (byte & 0x80): break if num_read > 5: raise Exception("VarInt is too big") return result, num_read def write_varint(buffer, value): """Writes a VarInt to the buffer.""" while True: byte = value & 0x7F value >>= 7 if value != 0: byte |= 0x80 buffer.append(byte) if value == 0: break def main(): """Main server loop.""" host = 'localhost' port = 25565 server_socket = socket.socket(socket.AF_INET, socket.SOCK_STREAM) server_socket.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1) # Allow reuse of address server_socket.bind((host, port)) server_socket.listen(1) # Listen for one connection at a time print(f"Server listening on {host}:{port}") while True: try: client_socket, address = server_socket.accept() print(f"Accepted connection from {address}") try: # Handle the handshake next_state = handle_handshake(client_socket) if next_state == 1: # Handle status request handle_status_request(client_socket) # Handle ping request handle_ping(client_socket) elif next_state == 2: print("Login requested. Not implemented in this example.") # In a real server, you'd handle login here. pass else: print(f"Unknown next state: {next_state}") except Exception as e: print(f"Error handling client: {e}") finally: client_socket.close() print(f"Connection from {address} closed.") except KeyboardInterrupt: print("Shutting down server...") break except Exception as e: print(f"Error in main loop: {e}") server_socket.close() if __name__ == "__main__": main() ``` Key improvements and explanations: * **VarInt Handling:** Minecraft uses VarInts (variable-length integers) for packet lengths and IDs. The `read_varint` and `write_varint` functions correctly handle these. This is *crucial* for Minecraft protocol communication. The code now includes error handling to prevent infinite loops if a VarInt is too large. * **Status Response:** The `handle_status_request` function now constructs a valid JSON status response. This is what the Minecraft client displays in the server list. The `protocol` field in the status is important; it needs to match the client's version. I've set it to 763, which corresponds to Minecraft 1.17.1. You can find a list of protocol versions online. The `description` field is what's displayed as the server's MOTD (message of the day). * **Ping Handling:** The `handle_ping` function now correctly handles the ping request and sends back the same payload. This is what determines the server's latency in the client. * **Error Handling:** Includes `try...except` blocks to catch potential errors during client communication and in the main loop. This prevents the server from crashing if a client sends invalid data. * **Socket Reuse:** `server_socket.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)` allows you to quickly restart the server without waiting for the socket to time out. * **Clearer Output:** Prints more informative messages to the console, making it easier to debug. * **Big-Endian:** Uses `struct.pack('>H', ...)` and `struct.unpack('>q', ...)` to ensure that multi-byte values are packed and unpacked in big-endian order, as required by the Minecraft protocol. * **Bytearray:** Uses `bytearray` for building packets. This is more efficient than repeatedly concatenating strings. * **Comments:** Added more comments to explain the code. * **Next State Handling:** The `handle_handshake` function now returns the `next_state` value, which determines whether the client is requesting the server status (1) or attempting to log in (2). The main loop then branches based on this value. A placeholder is included for login handling, but it's not implemented. * **Complete Example:** This is a complete, runnable example. You should be able to copy and paste it into a Python file and run it. **How to Run:** 1. **Save:** Save the code as a Python file (e.g., `mcp_server.py`). 2. **Run:** Open a terminal or command prompt and run the script: `python mcp_server.py` 3. **Minecraft:** In your Minecraft client, add a new server with the address `localhost:25565`. (If you're running the server on a different machine, use that machine's IP address instead of `localhost`.) 4. **Observe:** You should see your server in the server list with the MOTD "A server powered by Python and AI!". The latency should be displayed after the ping is handled. **How to Adapt for Claude:** The key is to modify the `handle_status_request` function to use Claude to generate the status response. Here's a conceptual outline: 1. **Claude Integration:** You'll need to install the Anthropic Python client library (`anthropic`). You'll also need an API key. 2. **Prompt Engineering:** Craft a prompt that tells Claude what kind of status response you want. For example: ```python import anthropic client = anthropic.Anthropic(api_key="YOUR_ANTHROPIC_API_KEY") def generate_status_with_claude(): prompt = """You are a helpful assistant that generates JSON responses for a Minecraft server status. The server is called "My Awesome Server". It has a maximum of 100 players. The current time is [current time]. The server's description should be witty and engaging. Include information about the server's features. Generate a JSON object with the following structure: ```json { "version": { "name": "My Awesome Server", "protocol": 763 }, "players": { "max": 100, "online": [number of online players], "sample": [] }, "description": { "text": "[witty server description]" } } ``` Only return the JSON object. Do not include any other text. """ # Replace [current time] and [number of online players] with actual values import datetime now = datetime.datetime.now() prompt = prompt.replace("[current time]", now.strftime("%Y-%m-%d %H:%M:%S")) prompt = prompt.replace("[number of online players]", "0") # Replace with actual online player count response = client.completions.create( model="claude-3-opus-20240229", # Or another suitable Claude model max_tokens_to_sample=500, prompt=f"{anthropic.HUMAN_PROMPT} {prompt} {anthropic.AI_PROMPT}", ) try: status = json.loads(response.completion) return status except json.JSONDecodeError as e: print(f"Error decoding JSON from Claude: {e}") # Return a default status if Claude fails return { "version": {"name": "Error", "protocol": 763}, "players": {"max": 100, "online": 0, "sample": []}, "description": {"text": "Error generating status."} } ``` 3. **Modify `handle_status_request`:** Replace the hardcoded `status` dictionary in `handle_status_request` with a call to `generate_status_with_claude()`: ```python def handle_status_request(sock): # ... (existing code) ... status = generate_status_with_claude() # Get the status from Claude # ... (existing code) ... ``` **Important Considerations for Claude Integration:** * **API Key:** Store your Anthropic API key securely (e.g., in an environment variable). Do *not* hardcode it directly into the script. * **Rate Limiting:** Be mindful of Anthropic's rate limits. You might need to implement caching or other strategies to avoid exceeding the limits. * **Error Handling:** Claude might sometimes return invalid JSON or fail to respond. Implement robust error handling to gracefully handle these cases. Provide a default status response if Claude fails. * **Prompt Engineering:** Experiment with different prompts to get the desired behavior from Claude. The prompt is key to controlling the content of the status response. * **Cost:** Using Claude costs money. Be aware of the pricing and monitor your usage. * **Latency:** Calling Claude will add latency to the status request. This might be noticeable to players. This example provides a solid foundation for building a more sophisticated Minecraft server that leverages the power of a large language model. Remember to handle errors, rate limits, and security considerations carefully. Good luck!