Cursor and Windsurf both support MCP natively — useful when you're building internal tools that read Amazon data, or debugging DataHawk integrations alongside your code.

Prerequisites

• An active DataHawk workspace with MCP enabled
• Cursor or Windsurf installed

Setup (Cursor)

1. Open Cursor → Settings → Features → MCP
2. Click Add new MCP server
3. Configure:
   • Name: datahawk
   • Type: HTTP
   • URL: https://mcp.datahawk.co/mcp
4. Authorize via OAuth

Setup (Windsurf)

1. Open Windsurf → Settings → MCP servers
2. Add new server with the same endpoint
3. Authorize

Typical use cases

• Building internal dashboards — ask the IDE's AI to scaffold a React component that pulls specific DataHawk metrics
• Debugging — "Why is this ASIN showing 0 sales in my local dev environment but has revenue in DataHawk?" — the AI can query DataHawk directly to confirm
• Data modeling — explore what's available in DataHawk before writing ETL code against the BigQuery export

Tips

• IDE-based AIs often skip the validation step to save tokens — if you see weird errors, ask explicitly: "Call datahawk_validate_query first before loading."
• Pair MCP with the API + BigQuery access for production: MCP for exploration and prototyping, direct BigQuery for scheduled jobs