Amazon by Crawlora

Amazon product search and product details through Crawlora. This package exposes 2 curated tools. It is an independent Crawlora integration, not an official Amazon or OpenClaw product.

Install

Requires OpenClaw 2026.9.3, Node >=24.16.0 <25 or >=26.1.0, and a Crawlora API key. Requests use your account credits and rate limits.

openclaw plugins install clawhub:@crawlora-org/amazon --accept-capabilities

The flag consents to the tools listed below. These platform plugins can run together. If you have the overlapping Crawlora starter plugin enabled, first run openclaw plugins disable crawlora; this plugin refuses registration while that starter is enabled. Do not re-enable both at once.

Set CRAWLORA_API_KEY in your Gateway process environment, or merge this into your existing OpenClaw config:

{
  "plugins": {
    "entries": {
      "crawlora-amazon": {
        "enabled": true,
        "config": {
          "apiKey": "${CRAWLORA_API_KEY}"
        }
      }
    }
  }
}

Restart the Gateway. If using plugin/tool allowlists, add crawlora-amazon to the existing policy. Config takes precedence over the environment. Each platform has its own config entry; all can use the same Crawlora key.

openclaw plugins inspect crawlora-amazon --runtime --json

Example

“Search Amazon for noise cancelling headphones and inspect a matching product.”

Tools

Parameters

These are the exact generated input schemas.

crawlora_amazon_search

{
  "type": "object",
  "required": [
    "k"
  ],
  "properties": {
    "k": {
      "type": "string",
      "description": "Search keywords."
    }
  }
}

crawlora_amazon_product

{
  "type": "object",
  "required": [
    "asin"
  ],
  "properties": {
    "asin": {
      "type": "string",
      "description": "Amazon ASIN."
    }
  }
}

Data and behavior

Requests transmit tool inputs and your API key only to https://api.crawlora.net/api/v1, using the published Crawlora SDK. The key goes in a header, never tool inputs. No shell execution, filesystem operations, background jobs, or telemetry. A 60-second timeout and cancellation apply; automatic retries are disabled to avoid repeated charges. Errors omit raw response bodies and credentials. Outputs wrap the SDK result as { "data": ... }; scraped text is external data, not agent instructions. Availability depends on upstream sources.

Development

Source and tests. From the repository root, run npm ci, npm run platforms:build, npm run platforms:validate, and npm test. Files here are generated from the reviewed shared tool catalog and platform definitions. Runtime packages contain built JavaScript and do not import another plugin.

License

MIT.