Managing MCP Tools
Learn how Links dynamically generates AI tools from your connected Data Sources.
Unlike traditional AI integrations that offer a hardcoded list of static tools, the Links MCP Server provides dynamic tools. Every time you connect a new operational database, spreadsheet, or SaaS platform to your workspace, the system automatically generates a custom MCP tool for it.
This means your AI agents (whether that is the internal Links AI (K2) or an external client like Claude Desktop) can instantly read, search, and analyze your live financial data using the exact same connection architecture.
How to Create an MCP Tool
Because MCP tools are driven entirely by your raw data, creating a new tool is the exact same process as creating a Data Source.
The Golden Rule of Links MCP
Data Source + MCP Toggle = AI Tool. To expose any dataset to your AI agents, you simply navigate to the Data Sources dashboard, connect your system, and ensure the AI Enablement (MCP) toggle is set to Active.
Step-by-Step Configuration
We have a comprehensive, step-by-step guide on how to securely connect external systems, upload CSVs, and configure the necessary AI metadata descriptions.
Using Your Tools Externally
Once your Data Source is configured with MCP enabled, it is immediately available to the internal Links AI (K2) assistant.
However, if you want to use these tools in your own local environment (like Claude Desktop, Cursor, or Postman), you must securely connect your local client to the Links MCP Server using your Bearer Token.