Connect View as Markdown

MCP: Claude, Cursor and others

Add the two checks as tools in any client that speaks the Model Context Protocol. Then you just ask: "is this token safe to buy?", and the agent calls the check itself before it trades.

What you need

The addresshttps://staging.lorglabs.com/mcp with the public key, to try it. https://api.lorglabs.com/mcp with a key of your own, for real use.
A keySent as a header: Authorization: Bearer lorg_…. To try it, the public key: lorg_test_wxXht1MyxKVyQwjyxN5JPHGfatXIZbL9 (10 checks a minute). For your own, write to info@lorglabs.com.
TransportStreamable HTTP, stateless. Listing the tools needs no key; calling one does.

Claude Code

One command:

claude mcp add --transport http lorg-check https://staging.lorglabs.com/mcp \
  --header "Authorization: Bearer lorg_test_wxXht1MyxKVyQwjyxN5JPHGfatXIZbL9"

Cursor

In .cursor/mcp.json (this project) or ~/.cursor/mcp.json (every project):

{
  "mcpServers": {
    "lorg-check": {
      "url": "https://staging.lorglabs.com/mcp",
      "headers": { "Authorization": "Bearer lorg_test_wxXht1MyxKVyQwjyxN5JPHGfatXIZbL9" }
    }
  }
}

VS Code

In .vscode/mcp.json:

{
  "servers": {
    "lorg-check": {
      "type": "http",
      "url": "https://staging.lorglabs.com/mcp",
      "headers": { "Authorization": "Bearer lorg_test_wxXht1MyxKVyQwjyxN5JPHGfatXIZbL9" }
    }
  }
}

Claude Desktop

Claude Desktop reaches a remote server through the small mcp-remote bridge (it needs Node.js). In claude_desktop_config.json (Settings, Developer, Edit config), then restart Claude:

{
  "mcpServers": {
    "lorg-check": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://staging.lorglabs.com/mcp", "--header", "Authorization:${AUTH_HEADER}"],
      "env": { "AUTH_HEADER": "Bearer lorg_test_wxXht1MyxKVyQwjyxN5JPHGfatXIZbL9" }
    }
  }
}

The key sits in env because Windows breaks arguments that contain a space

Any other client

Any MCP client that can reach a remote server over Streamable HTTP and send a header works: give it the address and the Authorization header above.

The tools

Tool	Use it before	Arguments
`check_token`	Buying or swapping into a token	`chain` (`base` or `ethereum`), `address`, and `amount_usd` if the size of the trade is known
`check_address`	Sending money to a wallet or contract	`chain`, `address`

Each returns the same signed answer as the API (the answer). The server also tells the model how to use them: read action, show the summary, and treat a tool error as not verified, never as permission to trade.

Try it

Once connected, ask your agent in plain words:

Is 0x4ed4E862860beD51a9570b96d89aF5E1B0Efefed on Base safe to buy? I want to put in $500.

It calls check_token and answers with the verdict and the reasons. A check made through MCP counts against the key like any other call.