MCP Support
Quanty AI supports the Model Context Protocol (MCP), an open standard that allows AI models to securely interact with external tools and data sources. This enables your local companions to read files, query databases, and execute custom scripts while keeping the control layer entirely on your machine.
The real power of AI is unlocked when it knows about your world. MCP provides the "nervous system" for Quanty AI, allowing it to move beyond a simple chat box into a functional creative agent.
How MCP Works
MCP creates a standardized bridge between "Hosts" (like Quanty AI) and "Servers" (tools that provide data).
- Secure: All tool execution can be either on permission or automatically executed.
- Local: Most MCP servers can run on your own machine.
- Extensible: You can write your own MCP servers. Not inside Quanty AI but you can use them then.
MCP Capability Layers
Allow your companion to read local markdown notes, code repositories, or PDFs.
The AI can perform actions like searching the web, checking your calendar, or running code snippets.
Using Custom Tools
In the "MCP" panel, you can add any MCP-compliant server. Once connected, your companion will automatically gain the ability to use those tools during conversation.
I love learning new skills! Once you connect a tool via MCP, I can help you organize your files or even summarize your local documents without you having to copy-paste anything.
[!TIP] Check out the official MCP documentation to explore available servers and learn how to build your own.