Actions3
Create
AI-generatedSummary
Submit YouTube video URLs to Weld's background scraping service to extract comments.
Inputs
- urls (required) — YouTube video URLs to scrape comments from — one per line or comma-separated. Must start with http:// or https:// and be from youtube.com or youtu.be. Split on newlines or commas, trimmed and filtered empty.
- jobName — Optional name for the scrape job, submitted to the Weld API as the job name.
- waitForCompletion — Whether to wait for the job to finish and return results. When true, the n8n worker is held while polling the Weld API; for long-running jobs, use a separate Get node on a schedule.
- pollInterval — How often (in seconds) to check if the job is done. Only applicable when waitForCompletion is true. Default: 10.
- maxWait — Maximum time (in seconds) to wait before timing out. Only applicable when waitForCompletion is true. Default: 300.
- resultLimit — Maximum number of results to fetch after job completion. Only applicable when waitForCompletion is true. Default: 1000.
Output shape
When waitForCompletion is false, the raw API response from creating the scrape job (containing jobId and status) is returned immediately. When true, an array of result items is returned after the job completes or times out; each item contains json with _jobId, _scraperId, and the comment data (or { data: commentData }). If the job fails, an error is thrown (or returned as { error: '...' } if continueOnFail is enabled).
The job creation POST goes to /api/jobs/create with body { scraperId, name, inputs: [{url} ] }. Polling checks /api/jobs/get?jobId=... then /api/jobs/results?jobId=...&limit=... on completion.
Examples
Example 1: Immediate return with job ID
urls: 'https://youtube.com/watch?v=abc123'; waitForCompletion: false; jobName: 'my scrape' Returns: { jobId: '...', status: 'processing' }
Example 2: Wait for results
urls: '...'; waitForCompletion: true; pollInterval: 5; maxWait: 120; resultLimit: 50 Returns: Array of comment objects extracted from the videos.