Skip to content
mcp/harbor
Cursor

modelcostsaver for Cursor

io.github.sachinuppal/modelcostsaver

Offline MCP: predict LLM call cost and pick the cheapest capable model. No keys, no network.

client:Cursor transport:stdio runtime:npm

Install modelcostsaver in Cursor

~/.cursor/mcp.json · windows: %USERPROFILE%\.cursor\mcp.json 

{
  "mcpServers": {
    "modelcostsaver": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "@workswarm/modelcostsaver"
      ]
    }
  }
}

Use .cursor/mcp.json in a project root instead to scope it to that project. Cursor merges both. Cursor docs

modelcostsaver in other clients

modelcostsaver MCP server