Brave Videos Scraper

Search public Brave Videos results and export normalized, API-ready records. The Actor supports one query or a batch of up to 20 queries, fair result allocation, source-backed filters, deterministic limits, and Apify Residential or custom proxies.

Features

• Single-query and multiple-query search modes
• Fair interleaving so one query cannot consume the entire batch limit
• Brave time filters: past day, week, month, or year
• Duration and platform post-filters using metadata present in Brave's response
• Relevance, newest, view-count, and duration sorting
• Safe Search, country, and language settings
• Stable IDs, duplicate removal, and omission of unavailable optional values
• Structured Brave hydration extraction with a DOM fallback
• Direct, Apify Proxy, and custom proxy operation
• Bounded retries, block-page detection, and clean no-results behavior

Input

Provide either query, queries, or both. Duplicate queries are removed.

Field	Description	Default
query	One Brave Videos query	—
queries	Up to 20 queries	—
maxItems	Maximum records across all queries	100
maxItemsPerQuery	Optional cap for each query	fair share of maxItems
country	Two-letter Brave country code	us
language	Interface and result language	en
safeSearch	strict, moderate, or off	moderate
timeRange	any, day, week, month, or year	any
duration	any, short, medium, or long	any
platform	any or a supported source platform	any
sortBy	relevance, newest, viewsDesc, durationAsc, or durationDesc	relevance
maxConcurrency	Parallel query pages, 1–5	2
maxRequestRetries	Temporary navigation retries, 0–3	1
proxyConfiguration	Apify Proxy or custom proxy configuration	Apify Proxy in Console

Example batch input:

{
  "queries": ["machine learning tutorial", "generative AI news"],
  "maxItems": 20,
  "maxItemsPerQuery": 10,
  "country": "us",
  "language": "en",
  "safeSearch": "moderate",
  "timeRange": "month",
  "sortBy": "newest",
  "proxyConfiguration": {
    "useApifyProxy": true,
    "apifyProxyGroups": ["RESIDENTIAL"],
    "apifyProxyCountry": "US"
  }
}

Output

Each row contains only values Brave supplied or values deterministically derived from them. Unavailable optional properties are omitted instead of being emitted as null or empty strings. Depending on the source result, records can include:

• stable ID and global/query/page positions
• title, description, video URL, domain, video ID, and embed URL
• Brave, original, and resized thumbnail details
• duration label and numeric seconds
• channel, author, creator, publisher, and platform
• views, publication time, relative time, and source fetch time
• tags and source classification flags
• exact search context, applied filters, provenance, and scrape timestamp

The dataset schema documents 71 possible fields and provides overview and full-details views. A typical result contains 44–58 populated fields; fields that the source does not expose are not fabricated.

Filter notes

• timeRange is sent to Brave and verified in the applied search URL.
• duration and platform are enforced after extraction. Records missing required source metadata are excluded while those filters are active.
• Brave Videos does not expose a verifiable resolution filter. Non-empty legacy resolution input is rejected instead of silently returning unfiltered results.
• Brave currently exposes one video result page without a reliable continuation contract. The Actor therefore returns the available page and never waits on invented pagination.
• GOOGLE_SERP proxy is not applicable because this Actor targets Brave, not Google Search.

Performance and resilience

The Actor uses one consistent Firefox profile per run, blocks nonessential image/media/font downloads, extracts structured page hydration where available, and processes independent queries concurrently. Temporary network failures receive bounded retries. Challenge, CAPTCHA, and block pages are never stored as dataset records.

Local run

npm ci
npx playwright install firefox
apify run --purge --input-file INPUT.json

Results are written to storage/datasets/default.

Limitations

Public search results vary by query, country, language, Safe Search preference, network location, and Brave's current index. Some sources omit duration, view count, author, or publication metadata. The Actor does not solve CAPTCHAs or bypass access controls.