Starknet MCP Server is a Model Context Protocol (MCP) server that exposes various Starknet RPC methods, enabling interaction with Starknet's blockchain. It allows users to retrieve block information, transaction details, and state updates, making it ideal for real-time data queries in an MCP-supported environment like Claude Desktop. Developed and maintained by Milan Cermak, the open-source project supports read-only methods such as `starknet_getBlockWithTxHashes` and `starknet_call`. To use it, clone the repository, install dependencies with Bun, and configure the `claude_desktop_config.json` file. The initial release version was v0.1.0, and while some limitations exist, it integrates seamlessly for querying Starknet's blockchain state.
mcp-screenshot is an MCP server designed to capture screenshots and perform OCR text recognition. It supports capturing different screen regions (left half, right half, full screen) and offers multiple output formats like JSON, Markdown, vertical, and horizontal layouts. The tool integrates two OCR engines: yomitoku for high-accuracy Japanese text and Tesseract.js for Japanese and English. Developed by kazuph, mcp-screenshot is ideal for automating tasks that require text extraction from specific screen areas, making it a versatile solution for multilingual environments and easy integration with tools like Claude Desktop. It can be installed via `npx -y @kazuph/mcp-screenshot` and is released under the MIT License.
Elasticsearch MCP Server is a Model Context Protocol (MCP) server implementation designed to interact with Elasticsearch. It simplifies Elasticsearch operations by providing tools for index, document, and cluster management. Users can list indices, retrieve mappings, search documents using Query DSL, and check cluster health and statistics. This tool is beneficial for developers, data engineers, and DevOps professionals, especially those integrating natural language commands via platforms like Claude Desktop. The server can be run locally or in a Docker environment and is open-source under the Apache License Version 2.0.
maven-mcp-server is an MCP (Model Context Protocol) server designed to check Maven dependency versions. It allows users to verify dependencies, fetch the latest versions from the Maven Central Repository, and supports full Maven coordinates. This tool is beneficial for developers working with Java-based projects using Maven, Gradle, SBT, or Mill. It ensures real-time access to repository data and is useful for automated dependency management and version checks. maven-mcp-server can be installed globally via npm, run directly with npx, or integrated into development environments like Claude Desktop through Smithery. It is particularly helpful during development, CI/CD pipeline integration, and dependency audits. The server includes tools for retrieving the latest versions and checking specific versions of dependencies, with robust error handling. maven-mcp-server is released under the MIT License, making it open-source and free for both personal and commercial use.
Remote-MCP Server is a type-safe and bidirectional solution for remote Model Context Protocol (MCP) communication. It enables a local MCP client to connect to a remote MCP server, facilitating centralized management of model contexts. Created to meet immediate needs for remote access, it serves as an interim solution until the official MCP support is released in Q1 2025. Developers and organizations requiring immediate remote MCP functionalities can benefit from this library. Remote-MCP Server can be deployed on platforms like Cloudflare Workers and Node.js, with integrations for Nest.js. Currently in active development, its initial release (v0.1.0) is available, and future updates are expected. Core features include type-safe communication, basic MCP commands, crash-safe handling, event subscription, and authentication middleware. Contributions are welcome, and the project is licensed under the MIT License.
mcp-jdbc is a server implementation within the Quarkus/Java Model Context Protocol (MCP) framework that enables seamless interaction between MCP-enabled applications and JDBC-compatible databases. It supports various database systems such as Postgres, MySQL, Oracle, and Sqlite, enhancing the data management capabilities of AI-driven applications like Claude Desktop. Ideal for developers building scalable AI solutions, mcp-jdbc simplifies data storage and retrieval. The project is part of the Quarkus MCP Servers and can be found on GitHub at https://github.com/quarkiverse/quarkus-mcp-servers. Developers can run the mcp-jdbc server using JBang and contribute to its development by adding new server modules.
mcp-tung-shing is a Model Context Protocol (MCP) service that provides detailed Chinese Traditional Almanac (Tung Shing) information. It offers features such as auspicious and inauspicious times, suitable and unsuitable activities, and conversions between Gregorian and Lunar calendars. This service is ideal for event planning and understanding cultural practices, supporting modern development workflows via NPX or CLI. Maintained by baranwang, the source code and documentation are available on GitHub under the repository baranwang/mcp-tung-shing. The latest update was made on January 21, 2025. For integration, use either the NPX method or CLI, and the primary API endpoint is 'get-tung-shing', which accepts 'startDate' and 'days' parameters.
MCP Server Starter is a production-ready template for building Model Context Protocol (MCP) servers using TypeScript. It provides a clean, maintainable project structure and integrates tools like Bun for fast development, Biome for linting, and standard-version for automated version management. This template is ideal for developers aiming to build MCP servers efficiently, especially for environments like Claude Desktop. Maintained by StevenStavrakis on GitHub, it simplifies the process of creating, testing, and publishing MCP servers. Use MCP Server Starter when starting a new MCP server project to ensure robust structure and tooling.
Together AI Image MCP Server is an MCP (Multi-Command Protocol) server designed for generating images using Together AI's advanced image generation models. It integrates seamlessly with the Cline platform, offering features like resizing, maintaining aspect ratios, and producing high-quality JPEG outputs. This server simplifies the process of creating custom images by providing configurable parameters such as model selection, dimensions, and output directories, making it ideal for developers and designers. It is particularly beneficial for automating or enhancing image creation workflows, requiring basic Node.js knowledge and familiarity with Cline MCP settings. The server runs locally on macOS/Linux systems and is suitable for environments where command-line operations are preferred for automation tasks involving image generation. Consider using this server when you need to generate multiple customized images programmatically, especially for integrating AI-generated visuals into larger projects or applications.
Crew AI MCP Server is an advanced MCP server that leverages the CrewAI framework to provide robust AI agent and task management capabilities. It simplifies the creation and management of AI agents, allowing users to define roles, goals, and tasks, and execute workflows seamlessly. This tool is ideal for developers, researchers, and organizations looking to automate and streamline processes, especially in Python-based environments and with OpenAI APIs. Supported on macOS, Linux, and Windows (via Git Bash), it requires Python 3.8 or higher, jq, and VSCode with the Roo Cline extension. Use Crew AI MCP Server to manage complex workflows, assign and monitor tasks, and enhance research and data analysis efforts.
The Figma MCP Server is a Model Context Protocol (MCP) server that provides Cursor with access to Figma files. It simplifies and translates Figma API responses, offering only the most relevant layout and styling information, which enhances AI accuracy in generating code from design data. This tool allows developers to fetch and implement Figma design metadata directly into code, improving productivity and accuracy compared to traditional methods. Developers and teams using Cursor as an AI coding assistant can benefit from this integration, streamlining the design-to-code process. The Figma MCP Server can be used locally or integrated into development environments, connecting to Cursor via HTTP mode and working seamlessly with tools like npm, pnpm, yarn, and bunx. Setting it up is particularly useful during early UI development or when design updates are frequent.
MCP Tool Server is a project that showcases the implementation and usage of Anthropic's Model Context Protocol (MCP) with AWS Bedrock. It offers a client that interacts with MCP-enabled tools via AWS Bedrock’s runtime service, supporting both stdio and sse communication modes. The server simplifies model-tool interactions by handling format conversion, asynchronous communication, and structured logging. Developers and organizations using AWS Bedrock and MCP can benefit from this tool, especially for integrating and testing AI workflows. MCP Tool Server requires Python 3.10 or higher, an AWS account with Bedrock access, and the UV package manager. The source code is available on GitHub under davidshtian/MCP-on-AWS-Bedrock. Key features include seamless integration, tool format conversion, and structured logging. Contributions are welcome, and additional resources are available for further learning.