Naver Video Search Collector

Collect public Naver video-search results for one or more keywords and store normalized records in the Actor's default dataset. The Actor uses direct HTTP requests and does not require customer credentials, browser automation, or a proxy.

Input

{
  "keywords": ["python", "data engineering"],
  "maxItemsPerKeyword": 100,
  "sort": "relevance",
  "concurrency": 3
}

keywords accepts one non-empty string or a non-empty list of strings. maxItemsPerKeyword ranges from 1 to 1000. sort is relevance, date, or views; concurrency ranges from 1 to 10. Defaults are 100, relevance, and 3 respectively.

Output

Each dataset item contains title, url, thumbnail, duration (seconds), channel, official, uploadDate, viewCount, sourcePlatform, keyword, ranking, sort, and the UTC collectedAt timestamp. Optional source metadata may be null, and duplicate URLs are emitted only once per run.

Pricing behavior

The Actor uses pay-per-event charging: one apify-actor-start event is emitted when a run starts, and one apify-default-dataset-item event is emitted after each dataset item is successfully stored. Failed persistence is not charged. The configured monetary price is managed by the platform release configuration.

Limitations

Naver's server-rendered result markup and sort parameters are undocumented and may change. Results and optional metadata depend on what Naver exposes publicly at run time. Results are deduplicated across all supplied keywords, so the first keyword that finds a URL owns that output item.

Functional input contract

CollectorConfig.from_mapping accepts:

• keywords: one non-empty string or a non-empty list of strings;
• maxItemsPerKeyword: integer 1–1000 (default 100);
• sort: relevance, date, or views (default relevance);
• concurrency: integer 1–10 (default 3).

Functional output contract

Each record contains title, url, thumbnail, duration in seconds, channel, official, uploadDate, integer viewCount, sourcePlatform, keyword, one-based ranking, sort, and ISO-8601 collectedAt. Missing optional source fields are represented by null (and official defaults to false). URLs are deduplicated across the entire run.

The caller supplies asynchronous persist(record) and charge(event) adapters. The collector calls charge("apify-default-dataset-item") only after the corresponding persist succeeds. The platform adapter owns the run-start event.