Startpage Video Scraper

Collect dedicated Startpage Video results as rich, normalized records. The Actor supports one search or a fair batch of up to 20 searches, native session-preserving pagination, duration/view/date/channel/domain filters, Safe Search, sorting, and bounded browser controls.

Startpage's embedded video-* structured groups are the primary source. Validated first-party JSON responses and a semantic DOM fallback are retained for compatible page changes. Ads, ordinary web results, challenge pages, signed proxy links, and raw responses are never stored.

Example input

{
  "queries": ["machine learning tutorial", "javascript tutorial"],
  "maxItems": 50,
  "maxPages": 5,
  "sortBy": "views",
  "durationFilter": "medium",
  "minViews": 1000,
  "requireThumbnail": true,
  "safeSearch": "moderate",
  "language": "english",
  "maxConcurrency": 3,
  "proxyConfiguration": {
    "useApifyProxy": true,
    "apifyProxyGroups": ["RESIDENTIAL"],
    "apifyProxyCountry": "US"
  }
}

durationFilter supports any, short (under 4 minutes), medium (4–20 minutes), and long (over 20 minutes). Output sorting supports Startpage relevance, newest, most viewed, shortest, and longest. Optional publishedAfter and publishedBefore values use YYYY-MM-DD.

Use includeDomains, excludeDomains, includeChannels, and excludeChannels for source filters. region is an explicit output provenance label; select a proxy country for network geolocation.

GOOGLE_SERP is intentionally unsupported because it is an HTTP proxy for supported Google Search/Shopping requests, while this Actor visits Startpage through an HTTPS Firefox session. Use Apify Residential or custom proxies when direct access is unreliable.

Output

The 34-field contract can include stable and platform video IDs, positions/pages, title, canonical and embed URLs, domain, platform, channel, description, text and numeric duration, text and numeric views, ISO and source publication dates, direct thumbnail, source/extraction provenance, matched queries, requested filters, locale, and scrape time.

Unavailable optional descriptions or channel labels are omitted. The Actor never emits null, blank strings, empty arrays, fabricated fields, or signed anonymous-view URLs.

Reliability

Startpage Video pagination uses a page-generated POST form containing session state and a page token. The Actor validates and submits that native form instead of reconstructing fragile page= URLs. All requested pages and queries are buffered, normalized, filtered, deduplicated, fairly selected, and written atomically. A failed later page cannot leave a misleading partial-success dataset.

Sessions, persistent cookies, coherent Crawlee Firefox fingerprints, bounded retries, memory-aware concurrency, and optional Residential/custom proxies improve ordinary compatibility. CAPTCHA pages are detected and rejected; the Actor does not solve or bypass CAPTCHAs.

Local development

npm ci
npx playwright install firefox
npm test
apify run --purge --input-file qa-inputs/local-pagination.json
node validate-datasets.js

The QA fixtures cover single, fair batch, native pagination, filtered/sorted, clean-empty, low-concurrency, invalid-input, incompatible-proxy, and cloud Residential cases.