Connect MCP in a minute

MCP (Model Context Protocol) lets your AI assistant use real tools — read files, search the web, query a database — through a standard interface. Pair it with TDSPRO and one key powers both the model and the workflow.

The idea in one breath

Your editor (the MCP client) launches small MCP servers. Each server exposes tools the model can call. You declare them in a tiny JSON config — that is the whole setup.

Step 1 — add the config

Most clients read an mcpServers block. Here is a minimal example with a filesystem and a fetch tool:

{
  "mcpServers": {
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-filesystem", "."]
    },
    "fetch": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-fetch"]
    }
  }
}

Step 2 — point the model at TDSPRO

In the same client, set the OpenAI-compatible base URL and your key. Now the model behind your tools is any of TDSPRO’s — pick by alias:

Base URL:  https://api.tdspro.lol/v1
API Key:   sk-hub-your-key-here
Model:     smart   # or fast / code / vision / chat

Step 3 — verify it works

1. 1Restart your editor so it picks up the config.
2. 2Ask the assistant to “list files in the current folder”.
3. 3If it calls the filesystem tool, MCP is live. 🎉