# Changelog of Instagram Audience Scraper (`blackfalcondata/instagram-audience-scraper`) Actor

- **URL**: https://apify.com/blackfalcondata/instagram-audience-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/blackfalcondata/instagram-audience-scraper.md

## Changelog

All notable changes to this Actor are documented here.

### \[Unreleased] - 2026-10-01

#### Fixed

- Runs with detail lookups turned off now deliver the sampled accounts (what the audience sample itself carries: handle, profile link and whether the account is private) instead of ending with no results. Each delivered account is a normal billable row.

#### Changed

- Incremental runs: a change in follower, following or post counts or in engagement alone no longer counts as an update. An account is reported as updated when its handle, display name, bio, profile picture, verification or privacy changes. On your first run after this release nothing is re-delivered because of the change: existing accounts are quietly re-baselined and only real changes after that are reported.

### \[Unreleased] - 2026-09-30

#### Fixed

- A maximum cost per run is now honoured: the run plans its audience sample to what the limit pays for, stops when it is reached and finishes normally instead of being aborted. A limit that leaves nothing after the start fee ends the run with a clear message and no results.
- Accounts the limit cut off are no longer recorded as delivered in incremental state, so they are offered again on the next run.
- A run where the audited account's posts could be opened but yielded no audience at all now reports that the audience sample is short, instead of finishing as a normal audit.
- The end-of-run summary uses the current price of $0.002 per account.

### \[0.1.0] - 2026-09-02

First release.

#### Added

- Audience audit for any public Instagram account, with no login, cookies or session token.
- One scored row per sampled account: posts, follower/following ratio, bio, avatar and handle
  pattern, plus the flags behind the verdict, so every score can be re-derived from its own row.
- Engagement figures for the audited account: average likes, average comments and engagement rate
  over its recent posts.
- Accepts handles, @handles and profile URLs, and audits several accounts in one run.
- Incremental mode labels each account NEW, UPDATED or UNCHANGED against the previous run,
  comparing the account facts the score reads: followers, following, posts, privacy,
  verification, avatar and bio. A busy account's follower count moves between any two runs, so
  UPDATED is the normal result for one; the accounts that sit still — which is most of what a bot
  audit is looking for — report UNCHANGED.

#### Notes

- Private accounts return a single row explaining why their audience cannot be sampled.
- An account we could not read is never scored: a failed lookup is not evidence of a bot. It is
  also never billed — it is left out of the results entirely rather than charged for as an empty
  row, and that now covers the account you asked about as well as the sampled ones. A handle that
  genuinely does not exist is a different thing: that is an answer, so it is reported and charged.
  A run where nothing at all could be read ends as a failure carrying the reason, not as an empty
  success.
- An account whose handle no longer resolves is checked against its account id before being
  reported as gone, so a rename is followed rather than scored as a disappearance.
- Change detection needs the profile lookups: with them switched off a run reports no NEW,
  UPDATED or UNCHANGED at all, because it never measures the accounts and will not guess.