Jan 16
Loading video player...
Timecodes: 06:30 What is MCP in plain terms 10:35 Conntext7 MCP 13:39 NPX installation 14:28 FileSystem MCP server 15:00 PlayWright 15:36 AgentQL MCP 15:57 Create MCP in Python 22:12 Our Custom Python MCP Server Execution 24:31 DuckDuckGo MCP Server Links: Model Context Protocol servers ( One of the trustworthy ones since it's official): https://github.com/modelcontextprotocol/servers context7: https://github.com/upstash/context7 NPX installation: https://nodejs.org/en/download FileSystem MCP server: https://github.com/modelcontextprotocol/servers/tree/main/src/filesystem PlayWright: https://github.com/microsoft/playwright-mcp AgentQL MCP: https://docs.agentql.com/integrations/mcp DuckDuckGo-mcp-server: https://github.com/nickclyde/duckduckgo-mcp-server Description : Unlock the power of local LLMs with real-time automation using MCP (Model-Context Protocol). In this video, we go beyond quantization and safe tensor formats to explore why running LLMs locally matters—especially for financial markets, private workflows, and tool-based automation. You'll learn: - ✅ Why local LLMs outperform cloud APIs for privacy, speed, and control - 🛠️ What MCP is and how it connects LLMs to tools, APIs, and real-time data - 🧩 How to build your own MCP server in Python (with a toy example) - 📦 Which models to deploy locally for tool calling (7B–10B, q8 quantized) - 🧠 How to use LM Studio to install and manage MCP servers - ⚙️ Real-world MCPs like Quantext7, Playwright, AgentQL, and FileSystem - 💡 Tips for context length, model selection, and automation setup Whether you're building a stock market advisor, a gym trainer, or a browser automation agent—this video gives you the blueprint to create your own AI assistant with full local control.