📸 Instagram Tagged & Mentions Posts Scraper

Scrape Instagram tagged posts, a profile's own posts, or a best-effort mentions approximation — with full media, carousels, usertag coordinates, location, audio, branded-content data, and optional comments.

This actor talks to Instagram's private web API. It resolves a target's numeric user id from web_profile_info and reads the genuine usertags feed (/api/v1/usertags/{user_id}/feed/) — the real surface of posts that other accounts tagged the target in. It does not fabricate a "tagged" concept from the target's own timeline, and it has no hardcoded related-accounts list.

Modes

Mode	What it returns	Auth
tagged	Posts where other accounts tagged the target (the real usertags feed)	sessionid required
profile-posts	The target's own timeline posts	Public accounts work without login
mentions	⚠️ Best-effort approximation only — see below	sessionid recommended

⚠️ About mentions

Instagram has no global, queryable index of "public posts that @-mention account X." usertags is photo-tags only; the mentions inbox is scoped to the logged-in account and cannot be queried for a third party. This actor therefore implements mentions only as a clearly-labeled approximation: it scans the candidate accounts and/or hashtags you supply and keeps posts whose caption @-mentions or whose usertags include the target. It never fabricates a mentions feed. If you supply no candidates or hashtags, it returns nothing.

Authentication (sessionid)

The tagged feed and comments are login-walled. Provide your Instagram sessionid cookie in the Instagram sessionid Cookie field. Use a sticky residential proxy (the default) — the actor automatically pins one IP for the whole run when a session is supplied, because rotating IPs under one login cookie triggers checkpoints. If the session is expired or flagged, the tagged feed returns login_required and the actor reports it honestly and emits no rows — it never fakes data.

Media URLs expire

Instagram's image/video CDN URLs (imageVersions, videoVersions, displayUrl) are signed and expire within hours. Download the media during or shortly after the run; do not treat these URLs as permanent.

Output

One row per post (type: "post"). When comments are enabled, each comment/reply is an additional child row (type: "comment", isChild, parentPostId, parentCommentId) — comments are mirrored to a per-run comments-<runId> dataset and do not count against maxPosts (which caps parent posts only).

Post fields include: postId, shortCode, url, ownerUsername, takenAt (ISO), mediaType, productType, likeCount, commentCount (-1 = comments disabled), videoViewCount, caption, hashtags, mentions, taggedUsers (with x/y position, verified flag, profile pic), coauthors, location*, isPaidPartnership, sponsors (from sponsor_tags), musicTitle/musicArtist, full imageVersions/videoVersions arrays, carouselMedia (each child's media

• its own usertags), author enrichment (ownerFollowerCount, ownerBiography, …), and derived engagementRate, postAgeDays, aspectRatio, scrapedAt.

Genuinely unavailable values are emitted as null — never a faked 0 or a stand-in. The base actor's fabricated fields (viewer-context owner{} block, always-0 edge_* counts, always-[] sponsor_user, mentioned_user, hardcoded post_type, dead success) have been removed.

Filters

onlyPostsNewerThan (absolute or relative date), minLikes, minComments, mediaType (image/video/carousel), keywordInCaption. Posts are de-duplicated by post id across all targets.

Comments

Enable Include Comments to fetch comments via the correct numeric-media_pk endpoint (/api/v1/media/{media_pk}/comments/, plus child_comments/ for replies) with sessionid auth and pagination. maxCommentsPerPost caps parent comments; includeReplies adds child replies.

AI enrichment (optional, off by default)

Turn on Enable AI Enrichment and supply a model + key to add caption/comment sentiment and brand-mention detection. Supports Anthropic, OpenAI, Google Gemini, xAI Grok, DeepSeek, Perplexity, and Mistral (provider auto-detected from the model name). Best-effort: on any failure the AI fields stay null and scraping continues.

Pay-per-event

Billing uses two events: post_result (per top-level post) and comment_result (per comment/reply row). The event prices are configured by the actor owner in the Apify Console.

Keywords

instagram scraper, instagram tagged posts, instagram usertags, instagram mentions, instagram comments scraper, instagram carousel, instagram reels, instagram engagement, branded content, influencer marketing, social media data, instagram media, instagram location, instagram hashtags.