Coverage is deliberately uneven and follows what each network exposes rather than a common denominator. TikTok gets audience demographics, live-room status, Creative Center charts, sound lineage and Shop; Instagram gets highlight trays and reels-by-song; Twitch stops at profile, videos, schedule and clips. Handles are not always the key either — instagram_reels, tiktok_followers and twitter_followers take a numeric user_id, and twitter_comments, twitter_quotes and twitter_retweets take a post id, so a workflow usually resolves an identifier first. check_credits is the exception to all of that, and its own description tells the agent to call it before running paid calls.
A social-data retrieval server exposing 107 tools across TikTok, Instagram, YouTube, Twitch, LinkedIn, Facebook, Twitter/X, Reddit, Threads and Pinterest, plus Google web results and the TikTok, Meta, Google and LinkedIn ad libraries.
- Account readouts in the shape each network publishes them: tiktok_profile carries follower, following, total-like and video counts with verification status, instagram_profile adds post count and business status, youtube_channel returns subscribers and total views, and twitter_profile, threads_profile, twitch_profile, linkedin_profile, linkedin_company and reddit_subreddit_details do the same for their own networks.
- Post and video listings with the network's own paging token: tiktok_videos pages on max_cursor, instagram_posts on next_max_id, youtube_channel_videos and youtube_channel_shorts on continuationToken, twitter_user_tweets_all on cursor by numeric user_id, and facebook_profile_posts, facebook_profile_reels, twitch_user_videos and reddit_subreddit each take their own.
- Spoken content as text: tiktok_transcript, instagram_transcript, youtube_video_transcript, facebook_post_transcript, twitter_tweet_transcript and reddit_post_transcript return captions by URL, and tiktok_video_info can fetch metadata and transcript together.
- Comment threads to the reply level: tiktok_comments with tiktok_comment_replies, youtube_video_comments with youtube_video_comment_replies, facebook_post_comments with facebook_comment_replies, instagram_comments, reddit_post_comments, and twitter_comments alongside twitter_quotes and twitter_retweets.
- Ad transparency, platform by platform: tiktok_ad_library_search filters by industry, objective, format, language, region and engagement; facebook_ad_library_search, facebook_ad_library_company_ads and facebook_ad_library_search_companies cover Meta; google_ad_library_company_ads, google_ad_library_ad_details and google_ad_library_search_advertisers cover the Ad Transparency Center; linkedin_ad_library_search and linkedin_ad_library_ad_details cover LinkedIn.
- Discovery feeds rather than named accounts: tiktok_trending, tiktok_music_popular, tiktok_creators_popular, tiktok_videos_popular and tiktok_hashtags_popular read the Creative Center charts by period and country, and youtube_shorts_trending returns what is climbing on Shorts.
- Audio lineage where a network exposes it: tiktok_search_music, tiktok_music_details and tiktok_music_videos go from a keyword to a clipId to the videos using it, and instagram_reels_by_song does the same from an audio_id.
- Commerce listings: tiktok_shop_search, tiktok_shop_products, tiktok_shop_product_details and tiktok_shop_product_reviews cover TikTok Shop, and facebook_marketplace_location_search, facebook_marketplace_search and facebook_marketplace_item cover Marketplace by coordinates with price, condition and delivery filters.
- check_credits reports the balance left on the configured key, which the scraping calls draw down.
A SociaVault API key, supplied as SOCIAVAULT_API_KEY, with credit on it — the scraping calls are paid and check_credits is what reports the balance.
One command plus a key — npx -y sociavault-mcp, then supply credentials