Quickstart

Connect your first MCP client, run a verified query, and then move to production automation.

Prerequisites

• An Attribution.ai workspace with an active connected store
• Access to dashboard API settings (Connect Your AI)
• At least one recent synced order so query results are non-empty

Step 1: Open AI Connector Setup

1. Go to Connect Your AI in the dashboard
2. Create credentials for your integration
3. Use OAuth + PKCE for interactive clients whenever available

Step 2: Configure Your Client

Use the AI Connectors guide for platform-specific steps. The snippets below are the fastest way to configure direct MCP access.

{
  "mcpServers": {
    "attribution-ai": {
      "url": "https://attribution.ai/api/mcp/rpc",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Step 3: Validate Discovery

A healthy flow calls tools/list or discover first, then targeted tools like performance or trends.

Step 4: Add CLI Access (Optional)

Install and verify CLI access:

npm install -g attribution-ai

attribution auth login
attribution tools list
attribution data query --tool performance --args '{"start_date":"2026-01-01","end_date":"2026-01-31"}'

Step 5: API Smoke Test

curl -s -X POST https://attribution.ai/api/mcp/rpc \
 -H "Content-Type: application/json" \
 -H "Authorization: Bearer atai_your_key_here" \
 -d '{"jsonrpc":"2.0","method":"tools/list","id":1}'

A successful response returns a result.tools array. If you receive 401/429 responses, use Error Reference for exact fixes.

Next Steps

• Use the connector setup guide for ChatGPT, Claude, and Gemini
• Install the CLI for automation and agent workflows
• Browse all 16 tools available via MCP
• Learn about authentication and rate limits
• Use the API reference for direct HTTP access