Discover Awesome MCP Servers

CPersona

Persistent AI memory server with 3-layer hybrid search (vector + FTS5 + keyword), confidence scoring via Reciprocal Rank Fusion, episodic/profile memory, and 16 tools. Zero LLM dependency. Works standalone with Claude Desktop and Claude Code. MIT licensed.

Agent Context API MCP

MCP server for deterministic context preflight, token counting, and credential redaction for coding agents, with a free local tier up to 12,000 characters and paid HTTP fallback for larger inputs.

FlashCardMCP

Converts JSON/CSV Markdown content into interactive flashcard pages with multiple templates, PDF export, and voice support for language learning.

weather-mcp

A FastMCP server that provides current weather, forecasts, and rule-based umbrella/jacket recommendations via Open-Meteo, deployable to Databricks Agent Bricks.

Enterprise Support MCP Server

A prototype that exposes customer and order data to LLMs via the Model Context Protocol, enabling natural-language queries and multi-step tool calling.

AgentOps

MCP server that browses the web, checks conditions (price, text, stock), and takes autonomous actions like form filling, with n8n notifications and a dashboard.

agent-fleet

MCP server that enables Claude Code to delegate tasks to Codex for real-time collaborative code generation and execution.

health-mcp

Provides a unified MCP interface to a personal health database, enabling AI agents to sync lifting workouts from Hevy, log food intake against a catalog, and run read-only SQL queries to cross-reference training and nutrition data.

cypress-mcp-server

A Model Context Protocol (MCP) server that integrates Cypress with Cursor AI, enabling AI-powered test automation, debugging, and script generation for Cypress tests.

OpenCode Advisor MCP

An unofficial MCP server that lets Codex ask a read-only OpenCode advisor for a second review pass on Git changes.

Cortex XSIAM MCP Gateway

Enables governed agent access to Cortex XSIAM security operations APIs, including XQL log search, issues, cases, endpoints, and assets, with dataset authorization and optional identity-based access control.

speaches-mcp

An MCP server that exposes speech-to-text and text-to-speech capabilities using a local speaches instance, allowing AI assistants to transcribe audio and generate speech.

Verging Labs MCP server

Provides AI agents with honest benchmark rankings (Agentic Memory Index and Agentic Search Index) for AI tools, plus graded checks and telemetry for x402 endpoints.

astronomy-mcp-server

Provides offline astronomical computations for planet and moon positions, rise/set times, phases, eclipses, and seasons via MCP.

chrome-feedback

MCP server for interactive feedback collection via Chrome extension, enabling users to submit text, images, and page elements.

Figma To Code MCP

Transforms Figma design data into a compact, LLM-friendly format for code generation, reducing size by 99.5% while preserving UI-critical information.

MindMeister MCP Server

Connects Claude to the MindMeister API v2, enabling AI-powered mind map management including viewing, listing, exporting, and sharing maps directly from Claude.

Inkwell

Self-hosted MCP server giving AI agents persistent memory — Markdown source of truth, hybrid BM25+embedding search, typed graph relations.

AMFI MCP Server

Provides AI assistants with real-time Indian mutual fund NAV data from AMFI's official feed, requiring no API key.

perplexity-mcp-server

Enables interaction with Perplexity AI through MCP tools for chatting, searching, and retrieving documentation.

simplenote-mcp

A remote MCP connector that gives Claude.ai projects a persistent, cross-conversation file store, backed by Simplenote and deployed free on Cloudflare Workers.

MCP Python Interpreter with Docker

Enables executing Python code, managing environments and packages, and performing file operations within Docker containers.

Terminal MCP

Enables management of visible, interactive terminal sessions across platforms (macOS, Windows, Linux, WSL). Supports creating, executing commands, capturing output, and managing multiple terminal windows simultaneously.

Marketstack MCP Server

Exposes various Marketstack API endpoints as MCP tools, providing access to financial market data including EOD, intraday, splits, dividends, tickers, exchanges, and other financial information.

py-har-mcp

An MCP server for parsing and analyzing HAR files, enabling AI assistants to inspect network traffic with automatic redaction of sensitive authentication headers.

agent-vision-mcp

Enables non-vision LLMs to analyze images via any OpenAI-compatible vision API. Hardened against truncation, empty responses, and timeouts for reliable analysis.

health-mcp

A local-first MCP server that stores nutrition, biomarker, and wearable data in SQLite and exposes it as Model Context Protocol tools, enabling MCP-aware agents to log meals, query trends, and run correlations.

mcp_server

돌핀 MCP 클라이언트를 사용하여 샘플 MCP 서버를 구현하는 방법에 대한 안내입니다. **개요:** 이 가이드는 돌핀 MCP 클라이언트를 사용하여 간단한 MCP 서버를 구축하는 단계를 안내합니다. MCP (Minecraft Protocol)는 마인크래프트 클라이언트와 서버 간의 통신 프로토콜입니다. 이 샘플 서버는 기본적인 연결 수락, 핸드셰이크 처리, 그리고 간단한 상태 응답을 제공합니다. **필요 사항:** * **Python 3.6 이상:** 돌핀 MCP 클라이언트는 Python으로 작성되었습니다. * **돌핀 MCP 클라이언트 라이브러리:** `pip install dolphin-mcp`를 사용하여 설치합니다. * **네트워크 프로그래밍에 대한 기본적인 이해:** 소켓, 바이트 스트림 등에 대한 이해가 필요합니다. **구현 단계:** 1. **필요한 라이브러리 가져오기:** ```python import socket import struct import json from dolphin_mcp import packets from dolphin_mcp import utils ``` 2. **서버 설정:** ```python HOST = '127.0.0.1' # 로컬 호스트 PORT = 25565 # 마인크래프트 기본 포트 ``` 3. **소켓 생성 및 바인딩:** ```python server_socket = socket.socket(socket.AF_INET, socket.SOCK_STREAM) server_socket.bind((HOST, PORT)) server_socket.listen(1) # 한 번에 하나의 연결만 허용 print(f"서버가 {HOST}:{PORT}에서 시작되었습니다.") ``` 4. **클라이언트 연결 수락:** ```python client_socket, client_address = server_socket.accept() print(f"{client_address}에서 연결이 수락되었습니다.") ``` 5. **핸드셰이크 처리:** ```python # 패킷 길이 읽기 packet_length_bytes = client_socket.recv(1) packet_length = struct.unpack('>B', packet_length_bytes)[0] # 패킷 데이터 읽기 packet_data = client_socket.recv(packet_length) # 핸드셰이크 패킷 파싱 handshake_packet = packets.HandshakePacket.from_bytes(packet_data) print(f"프로토콜 버전: {handshake_packet.protocol_version}") print(f"서버 주소: {handshake_packet.server_address}") print(f"서버 포트: {handshake_packet.server_port}") print(f"다음 상태: {handshake_packet.next_state}") next_state = handshake_packet.next_state ``` 6. **상태 요청 처리 (상태 1):** ```python if next_state == 1: # 상태 요청 패킷 받기 (길이 1바이트, 내용 0x00) status_request_length_bytes = client_socket.recv(1) status_request_length = struct.unpack('>B', status_request_length_bytes)[0] status_request_data = client_socket.recv(status_request_length) # 상태 응답 생성 status_response = { "version": { "name": "샘플 서버", "protocol": handshake_packet.protocol_version }, "players": { "max": 100, "online": 0, "sample": [] }, "description": { "text": "돌핀 MCP 샘플 서버" } } status_response_json = json.dumps(status_response) status_response_bytes = status_response_json.encode('utf-8') status_response_length = len(status_response_bytes) # 패킷 길이와 데이터 전송 response_packet = struct.pack('>B', status_response_length) + status_response_bytes client_socket.sendall(response_packet) # Ping 요청 처리 (선택 사항) ping_request_length_bytes = client_socket.recv(1) ping_request_length = struct.unpack('>B', ping_request_length_bytes)[0] ping_request_data = client_socket.recv(ping_request_length) ping_payload = struct.unpack('>q', ping_request_data)[0] # long 타입으로 언팩 # Ping 응답 생성 ping_response = struct.pack('>q', ping_payload) ping_response_length = len(ping_response) # 패킷 길이와 데이터 전송 ping_response_packet = struct.pack('>B', ping_response_length) + ping_response client_socket.sendall(ping_response_packet) ``` 7. **로그인 처리 (상태 2 - 구현하지 않음):** ```python elif next_state == 2: print("로그인 요청은 아직 구현되지 않았습니다.") # TODO: 로그인 처리 구현 ``` 8. **연결 종료:** ```python client_socket.close() server_socket.close() print("연결이 종료되었습니다.") ``` **전체 코드 예시:** ```python import socket import struct import json from dolphin_mcp import packets from dolphin_mcp import utils HOST = '127.0.0.1' PORT = 25565 server_socket = socket.socket(socket.AF_INET, socket.SOCK_STREAM) server_socket.bind((HOST, PORT)) server_socket.listen(1) print(f"서버가 {HOST}:{PORT}에서 시작되었습니다.") client_socket, client_address = server_socket.accept() print(f"{client_address}에서 연결이 수락되었습니다.") # 핸드셰이크 packet_length_bytes = client_socket.recv(1) packet_length = struct.unpack('>B', packet_length_bytes)[0] packet_data = client_socket.recv(packet_length) handshake_packet = packets.HandshakePacket.from_bytes(packet_data) print(f"프로토콜 버전: {handshake_packet.protocol_version}") print(f"서버 주소: {handshake_packet.server_address}") print(f"서버 포트: {handshake_packet.server_port}") print(f"다음 상태: {handshake_packet.next_state}") next_state = handshake_packet.next_state if next_state == 1: # 상태 요청 status_request_length_bytes = client_socket.recv(1) status_request_length = struct.unpack('>B', status_request_length_bytes)[0] status_request_data = client_socket.recv(status_request_length) status_response = { "version": { "name": "샘플 서버", "protocol": handshake_packet.protocol_version }, "players": { "max": 100, "online": 0, "sample": [] }, "description": { "text": "돌핀 MCP 샘플 서버" } } status_response_json = json.dumps(status_response) status_response_bytes = status_response_json.encode('utf-8') status_response_length = len(status_response_bytes) response_packet = struct.pack('>B', status_response_length) + status_response_bytes client_socket.sendall(response_packet) # Ping 요청 ping_request_length_bytes = client_socket.recv(1) ping_request_length = struct.unpack('>B', ping_request_length_bytes)[0] ping_request_data = client_socket.recv(ping_request_length) ping_payload = struct.unpack('>q', ping_request_data)[0] ping_response = struct.pack('>q', ping_payload) ping_response_length = len(ping_response) ping_response_packet = struct.pack('>B', ping_response_length) + ping_response client_socket.sendall(ping_response_packet) elif next_state == 2: print("로그인 요청은 아직 구현되지 않았습니다.") client_socket.close() server_socket.close() print("연결이 종료되었습니다.") ``` **실행 방법:** 1. 위 코드를 `mcp_server.py`와 같은 파일로 저장합니다. 2. 터미널에서 `python mcp_server.py`를 실행합니다. 3. 마인크래프트 클라이언트를 실행하고, 서버 주소를 `127.0.0.1:25565`로 설정하여 연결합니다. **주의 사항:** * 이 코드는 매우 기본적인 샘플 서버이며, 실제 마인크래프트 서버의 모든 기능을 제공하지 않습니다. * 에러 처리, 보안, 성능 최적화 등은 추가적으로 구현해야 합니다. * 돌핀 MCP 클라이언트 라이브러리의 최신 문서를 참고하여 더 많은 기능을 활용할 수 있습니다. * `struct.pack`과 `struct.unpack`을 사용하여 바이트를 패킹/언패킹할 때, 바이트 순서(endianness)와 데이터 타입을 정확하게 지정해야 합니다. 위 예시에서는 `>` (big-endian)과 `B` (unsigned byte), `q` (long long)를 사용했습니다. **다음 단계:** * 로그인 처리 구현 * 채팅 기능 구현 * 플레이어 위치 정보 처리 * 청크 데이터 전송 * 더 많은 마인크래프트 프로토콜 패킷 처리 이 가이드가 돌핀 MCP 클라이언트를 사용하여 마인크래프트 서버를 구축하는 데 도움이 되기를 바랍니다. 궁금한 점이 있으면 언제든지 질문해주세요.

velesdb-memory

Local-first agent-memory MCP server with a why() tool: recall a fact together with its connected subgraph (multi-hop), so linked memories surface even when they share no words with the query. remember/recall/relate/forget/why over one fused vector + graph + columnar engine a single offline Rust binary.