Jobs

Batch, deep, and scheduled crawls use POST /v1/jobs. Download JSONL / CSV / Parquet, or sink to your bucket (AssumeRole supported).

Create

Console wizard: upload / map → estimate → delivery → submit. Or SDK:

await client.jobs.create({
  endpoint: "xiaohongshu.note.detail",
  params: [{ note_id: "…" }],
});

Large inputs: stream via params_url.

Priority

bulk / standard / priority affect queue and price. Failed items can be retried.

Marketing: /jobs.