Through MCP, your own AI can request data and use functions outside the chat.
How it works
An AI application talks through MCP to a server that offers data or functions, for example a database, a calendar or a calculation model. Its makers compare it to a USB-C port for AI: one standard way to connect.
Why it matters
With MCP you no longer need to paste exports into a chat. But a connection alone says nothing about quality and security: what the AI sees, how it is defined and who may do what is arranged on the server side.
MCP in Dashly
With Dashly Access you connect Claude, ChatGPT, Copilot or your own agents to Dashly through MCP. From Knowledge onwards your AI calculates with the same fixed formulas as your dashboards, with permissions per user and per agent, and every question is logged.