@dropvalidate/mcp

MCP server for DropValidate — get PASS/FAIL product verdicts (scored out of 10, backed by ad-library, Reddit sentiment, and demand-trend receipts) inside Claude, Cursor, or any MCP client.

It's a thin wrapper over the DropValidate account API. Every tool authenticates with your own dv_live_… key and scopes to your account.

Setup

1. Create an API key at https://app.dropvalidate.com/settings/api-keys (use a read + write key if you want the agent to run scans).
2. Add the server to your client config:

{
  "mcpServers": {
    "dropvalidate": {
      "command": "npx",
      "args": ["-y", "@dropvalidate/mcp"],
      "env": { "DROPVALIDATE_API_KEY": "dv_live_your_key_here" }
    }
  }
}

• Claude Desktop: Settings → Developer → Edit Config.
• Cursor: Settings → MCP → Add.

That's it — no separate login. (Optional DROPVALIDATE_API_BASE env var overrides the API base URL.)

Tools

Read (free): get_account, get_credits, get_credit_packs, list_validations, get_validation, get_validation_status, list_winners, get_winners_search, list_strategies, get_strategy, list_portfolios.

Write (spend credits — each requires confirm: true):

• validate_product — run a PASS/FAIL scan (5 credits); waits and returns the finished verdict by default.
• unlock_category — reveal the full report on a single-category scan (5 credits).
• run_winners_search — find products in a niche that PASS (10 credits).

Write tools never spend credits unless you pass confirm: true, so an agent can't burn credits on its own.

Example

"Validate this product: https://www.aliexpress.com/item/…"

The agent calls validate_product (with confirm: true), waits ~30s, and returns the verdict — PASS/FAIL, score out of 10, and the receipts behind it.

Buying credits

Credit purchases are in-app only (/credits) — the API and this MCP never move money.