Core FAQs

Question 1

How can I build my own MCP tool for Core?

Accepted Answer

You can easily build your own MCP tool using Python and the CoreMCP library. Simply decorate your tool functions with `@mcp.tool()` and place them in the `mcp_tools` folder for auto-registration.

Question 2

Is Core secure?

Accepted Answer

Yes, Core is designed to be local-first and secure. It runs offline, offers token-based authorization, and supports IPC or WebSocket routing.

Question 3

What are the key features of Core?

Accepted Answer

Core provides a single entry point (MCP Gateway) for all local tools, a dynamic web-based UI, Plugin SDKs for different IDEs and CLIs, Session Management for persistent workflows, and Dynamic Tool Registration for easily adding new MCP modules.

Question 4

What is Core and what does it do?

Accepted Answer

Core is a central hub for managing and orchestrating local AI tools using a unified Model Context Protocol (MCP) gateway. It allows you to connect and manage your AI tools across different environments, like IDEs, terminals, and custom AI assistants.

Question 5

What are Model Context Protocols (MCPs)?

Accepted Answer

MCPs define a standard way for different AI tools to communicate and share context. Core uses MCPs to provide a consistent and extensible system for managing your local AI tools.

Core

关于

主要功能

使用案例