ANDRU
Docs menu

Quickstart

From API key to first answer

About five minutes. You need Node.js 18 or later for MCP; A2A needs only an HTTP client.

1. Get an API key

  1. 1Sign in to Andru (a free account is enough) and open Settings › Developer.
  2. 2Create a key. It starts with sk_live_ and is shown once, so copy it somewhere safe.

2a. Claude Desktop or Cursor (MCP)

Add Andru to your client's MCP config. In Claude Desktop that is claude_desktop_config.json; in Cursor, ~/.cursor/mcp.json. Restart the client and Andru's tools appear.

claude_desktop_config.json
{
  "mcpServers": {
    "andru": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-server-andru-intelligence"
      ],
      "env": {
        "ANDRU_API_KEY": "sk_live_your_key_here"
      }
    }
  }
}

Using Claude Code instead:

Terminal
claude mcp add andru -e ANDRU_API_KEY=sk_live_your_key_here -- npx -y mcp-server-andru-intelligence

Then ask, for example, “Score Datadog against my ICP.” The package is mcp-server-andru-intelligence on npm.

2b. Your own agent (A2A)

Send a task naming a skill, the tool and its arguments. Lookups like this one are free.

Terminal
curl https://api.andru-ai.com/api/a2a/tasks \
  -H "X-API-Key: sk_live_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "id": "1",
    "method": "tasks/send",
    "params": {
      "skillId": "buyer-understanding",
      "message": {
        "role": "user",
        "parts": [{
          "kind": "data",
          "data": {
            "tool": "get_icp_fit_score",
            "args": {
              "companyName": "Datadog",
              "productDescription": "Observability platform for cloud infrastructure teams",
              "targetRole": "VP Engineering"
            }
          }
        }]
      }
    }
  }'

tasks/send waits for the answer and replies with a taskId, a status and the output. For long tasks, send with tasks/sendSubscribe instead and collect the result as described in Send a task.