Pinterest Following Scraper Actor

Apify actor that collects the following list of Pinterest profiles from usernames, profile URLs, or keywords. It mirrors the reference output in CandyW8119_following.json, streams data to the Apify dataset in real time, and follows a strict proxy fallback chain (direct → datacenter → residential).

Why this actor

• Built for Pinterest following discovery with pagination support
• Works with usernames, profile links, or keyword-like handles (bulk input)
• Live dataset writes and summary saved to key-value store
• Resilient anti-blocking: retries, random delays, and proxy fallbacks
• Minimal setup: just provide _pinterest_sess cookie and inputs

Key features

• Bulk inputs via sources (URLs, @usernames, keywords)
• Clean output fields: thumbnails, profile URL, follower/board counts, verification flags, etc.
• Proxy strategy: start with no proxy, then auto-switch to datacenter, then to residential and stick there if blocking is detected; all transitions are logged
• Retry with exponential backoff and 1–2s jitter between pages
• Async aiohttp requests for efficient pagination

Input

Example (Apify console JSON):

{
  "sources": [
    { "url": "https://www.pinterest.com/CandyW8119/" },
    { "url": "@anotheruser" }
  ],
  "pinterestSess": "YOUR__pinterest_sess_COOKIE",
  "sortOrder": "default",
  "maxComments": 120,
  "proxyConfiguration": {
    "useApifyProxy": false
  }
}

Fields

• sources (array, required): URLs/usernames/keywords (bulk supported).
• pinterestSess (string, required): _pinterest_sess cookie value.
• sortOrder (string): default, recent, or relevance (Pinterest may ignore).
• maxComments (integer): Max following records per input (1–1000).
• proxyConfiguration (object): Default is no proxy. On block → datacenter; if blocked again → residential (sticky). Groups honored when provided.

Output

• Data is pushed live to the default dataset. Each item matches the reference shape (sample from CandyW8119_following.json):
  • serachUsername, username, full_name
  • profileUrl, image_medium_url, image_large_url, pin_thumbnail_urls
  • follower_count, pin_count, board_count
  • Flags: is_verified_merchant, explicitly_followed_by_me, domain_verified, is_private_profile
  • Raw recent_pin_images for context
• Run summary is stored in key-value store under SUMMARY.

How it works

1. Normalize each source to a username.
2. Start direct requests with _pinterest_sess cookie.
3. Paginate /resource/UserFollowingResource/get/ with page size up to 50.
4. On block-like status (401/403/429/503): switch to datacenter; on further block: switch to residential and keep it. All changes are logged.
5. Retry each page up to 3 times with backoff; add 1–2s jitter between pages.
6. Transform each user and push immediately to the dataset; continue even if some inputs fail.
7. Save summary (processed inputs, total users saved, proxy mode/groups).

Running locally

$python -m src

Provide input via the Apify key-value store INPUT.json or environment when running on the platform. Ensure your _pinterest_sess cookie is valid.

Best practices

• Use a fresh Pinterest session cookie.
• Keep maxComments reasonable (e.g., <= 500) to reduce block risk.
• Allow proxy fallback; residential will stick after a block cascade.
• Monitor logs for proxy switches and retry events.

Support and notes

• Data is collected only from public endpoints.
• Respect Pinterest terms and local regulations when using the data.