Latest insights, tutorials, and updates from the Model Context Protocol team
kilocode is an open-source TypeScript MCP Client designed as an AI coding assistant. It helps developers plan, build, and fix code through features like code generation, task automation, and automated refactoring. This client extends capabilities by integrating with various MCP servers via its marketplace.
The ai-engineering-from-scratch MCP Client is a Python-based educational and development resource, offering 503 lessons across 20 phases to guide developers through building and shipping AI systems. It serves as a comprehensive reference for AI engineering, providing structured learning to integrate with MCP ecosystems.
haystack is an open-source Python MCP Client for building context-engineered LLM applications. It provides explicit control over retrieval, routing, memory, and generation, ideal for developers building scalable agents and RAG systems.
cc-switch is a Rust-based desktop MCP Client that unifies management for Claude Code, Codex, OpenCode, OpenClaw, and Gemini CLI. It provides a single panel for MCP server configuration, prompt synchronization, and skill installation across these diverse AI tools. Developers can centralize their AI assistant workflows, ensuring consistent context and capabilities.
Douyin_TikTok_Download_API is a Python-based, self-hosted MCP Server for scraping and downloading Douyin and TikTok content without watermarks. It provides a REST API, CLI, and web console, solving the problem of reliable, self-managed media archiving for developers.
cognee is an MCP Server providing a memory layer for AI applications and agents. It uses scalable ECL pipelines to interconnect and retrieve diverse data types, aiming to reduce hallucinations and development effort. Developers building AI systems needing dynamic, context-aware memory should consider cognee.