Instagram Hashtag Scraper - Top & Recent Posts

Extracts hashtag stats, top posts, and recent posts from Instagram. Fully cookieless by default; an optional session cookie improves reliability and unlocks reels.

Why this exists

The most-used Instagram Hashtag Scraper on Apify (57,685 users) sits at 2.98★ — the worst rating in the platform's top 20 — and its README reuses generic troubleshooting boilerplate word-for-word across unrelated actors from the same publisher, a strong signal of a template-cloned, unmaintained listing rather than an actively fixed one.

IMPORTANT — read before running

Unlike a documented public API (there is no official one for this), this actor talks to Instagram's internal, undocumented endpoints. That means:

• I could not test this against live Instagram from the build environment (no proxy access, Instagram not reachable). The parsing logic is built defensively — it searches responses for recognizable media objects rather than assuming one fixed JSON path — and is unit-tested against realistic sample shapes, but the first real run against live Instagram is the real test, the same way the SAM.gov actor's first live run was.
• A residential proxy is required. Instagram blocks datacenter IPs almost immediately. Use Apify's RESIDENTIAL proxy group (default in the input schema).
• If a run logs "No recognizable media objects found" on a hashtag known to have posts, that means Instagram changed its response shape — check the _raw field kept on every dataset item to see the actual response and adjust findMediaLikeObjects in src/ig-client.js accordingly, rather than assuming the hashtag is empty.

Setup

1. In the input, provide one or more hashtags (no #).
2. Leave sessionCookie blank for cookieless operation (posts/photos only), or paste your Instagram sessionid cookie for better reliability and to include reels.
3. Proxy defaults to Apify's RESIDENTIAL group — keep this on.

Reliability behavior

• Retries 429/5xx with exponential backoff; logs a clear note that repeated 429s usually mean the proxy IP needs rotating.
• A single hashtag failing (not found, or genuinely blocked) does not stop the rest of the batch from processing.
• If every hashtag in a run fails, the run fails loudly instead of quietly producing an empty dataset — that pattern almost always means a systemic problem (blocked proxy, changed endpoint), not coincidence.

Output

Two record types in one dataset, distinguished by recordType:

• hashtagStats: mediaCount (total public posts using the hashtag)
• post: postSource (top/recent), caption, like/comment/play counts, owner, media URLs, and a direct postUrl link.