Facebook Groups Search Scraper Plus

Discover and extract public Facebook group data from search keywords (via DuckDuckGo) and/or direct group URLs. This facebook groups scraper / facebook group finder returns numeric, sortable group metadata — member counts, posting activity, descriptions, and derived analytics — for community discovery, market research, and audience analysis, without login.

What it does

• Keyword discovery — turns plain keywords ("bmw", "photography") into Facebook group URLs using DuckDuckGo, then scrapes each group's public About page.
• Direct URLs — paste full facebook.com/groups/... links to scrape them directly. Mix keywords and URLs in one run.
• Numeric fields — memberCount, postsLastDay, postsLastMonth are parsed to integers so you can sort and filter by group size and activity.
• Descriptions — the group's og:description is harvested (coverage varies by IP/region; unavailable values are null, never faked).
• Filters — minMembers/maxMembers, privacy, minPostsPerDay, include/exclude keywords, and cross-keyword de-duplication.
• Derived analytics — activityLevel, postsPerWeek, engagementDensity, sizeBucket, membersPercentileRank.
• Optional AI — classify each group's description into category, topics, language, and audience via a multi-provider LLM (Anthropic/OpenAI/Google/xAI/DeepSeek/Perplexity/Mistral). Runs only when a description is present and a key is supplied.

Output fields (one row per group)

Field	Description
query	Keyword the group was found under, or "direct_urls"
numericGroupId / id	Facebook numeric group ID when detectable, else null
name	Group name (from og:title)
url	Canonical group URL
vanitySlug	Non-numeric slug from the URL, if any
profilePictureUri	Group image URL (og:image)
description	Group description (og:description), else null
visibility / isPrivate / isPublic	Public/Private when detectable, else null
memberInfo	Raw member text (e.g. "22,345 total members")
memberCount	Integer member count parsed from memberInfo
postsLastDay / postsLastMonth	Integer posting counts
postsPerWeek / engagementDensity	Derived activity metrics
activityLevel	very_active / active / low / dormant
sizeBucket	small / medium / large / huge
membersPercentileRank	0–100 percentile of memberCount within the run
resultTitle / resultSnippet / rankPosition	DuckDuckGo result title, snippet, and rank
aiCategory / aiTopics / aiLanguage / aiAudience	Present only when AI classification is enabled
scrapedAt	ISO 8601 timestamp of the run

Values that cannot be detected on a logged-out page are returned as null. Rows are only saved when a group name is found.

Input

{
  "startUrls": ["bmw", "tesla", "https://www.facebook.com/groups/123456789/"],
  "maxItems": 50,
  "minMembers": 1000,
  "excludeKeywords": ["marketplace"],
  "dedupe": true,
  "proxyConfiguration": { "useApifyProxy": false }
}

• startUrls (array, required) — keywords and/or group URLs.
• maxItems (integer) — max groups per keyword (1–1000). Direct URLs are always all scraped.
• maxTotalItems (integer) — optional hard cap across all keywords (0 = unlimited).
• minMembers / maxMembers / privacy / minPostsPerDay / keywords / excludeKeywords — filters.
• dedupe (boolean) — remove groups found under multiple keywords.
• aiEnhancement / aiModel / aiApiKey — optional LLM classification.
• proxyConfiguration (object) — optional; auto-escalates to datacenter then residential if blocked.

How it works

1. Discovery — for each keyword, query DuckDuckGo's internal JSON endpoint for "{keyword} facebook group" and harvest facebook.com/groups/... links plus each result's title, snippet, and rank.
2. Enrichment — fetch each group's logged-out /about page and parse OpenGraph meta tags and the embedded relay JSON.
3. Shaping — de-duplicate, apply filters, compute derived analytics and percentile rank, optionally classify with AI, then push to the dataset.

Notes and limits

• Discovery coverage is bounded by DuckDuckGo's index. Description/visibility availability on logged-out /about pages varies by IP and region; missing values are null.
• The actor collects only publicly accessible group metadata; it does not log in or scrape member lists. Follow applicable data-protection laws and platform terms.
• Export the dataset as JSON/CSV from the Apify console or fetch it via the Apify API.