YouTube by Crawlora

YouTube video transcripts through Crawlora. This package exposes 1 curated tools. It is an independent Crawlora integration, not an official YouTube 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/youtube --accept-capabilities

The flag consents to the tools listed below. These four 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-youtube": {
        "enabled": true,
        "config": {
          "apiKey": "${CRAWLORA_API_KEY}"
        }
      }
    }
  }
}

Restart the Gateway. If using plugin/tool allowlists, add crawlora-youtube 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-youtube --runtime --json

Example

“Get the transcript of YouTube video dQw4w9WgXcQ.”

Tools

Parameters

These are the exact generated input schemas. SEC company-specific tools additionally require ticker or cik.

crawlora_youtube_transcript

{
  "type": "object",
  "required": [
    "id"
  ],
  "properties": {
    "id": {
      "type": "string",
      "description": "YouTube video id."
    }
  }
}

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.