# Changelog of Google AI Overview Tracker 🔍 (is your brand cited?) (`tagadanar/google-ai-overview-tracker`) Actor

- **URL**: https://apify.com/tagadanar/google-ai-overview-tracker/changelog.md
- **Full Actor documentation**: https://apify.com/tagadanar/google-ai-overview-tracker.md

## Changelog

### 2026-09-14

- Pricing is now all-in: $0.03 per keyword checked plus a $0.005 run start
  fee, with the Google SERP proxy and compute included (in effect since
  September 5). There is no separate platform usage bill any more.

### 2026-09-08

- **A run can no longer spend more time than it has.** The retry ladder booked
  its worst case up front, under a loop with no ceiling, so a slow patch on the
  source could push the run past the platform timeout — and a run killed that
  way loses its own closing message entirely.
- Every request now plans against the time actually left, and a backoff or a
  politeness pause that would outlive the run is not slept.
- A run that hits its time limit says so and names exactly what it never
  reached, so an unchecked item is never reported as an empty one.

### 0.1.6

The control keyword is now checked lazily: a keyword that came back with an AI
Overview proves detection by itself, so the extra control search only runs when
at least one keyword needs a "no AI Overview" answer proven. Runs whose
keywords all show overviews get one search cheaper and faster.

A run that Google refuses outright now stops after the first two refused
searches instead of attempting the whole list; nothing is charged either way.

Announced pricing change: from September 5, 2026 the actor is all-in — $0.03
per keyword checked with platform usage included, replacing $0.008 per keyword
plus platform usage on your own account.

### 0.1.0

First release. Checks a list of keywords against Google and reports, per keyword,
whether an AI Overview appeared, its full text, the pages it cited, and whether
your brands or domains are cited in it.

Keywords that could not be checked come back marked `blocked` or `unknown` with
the reason, and are not charged. Every run also checks a control keyword first,
so a "no AI Overview" answer is only reported once the run has shown it can see
AI Overviews at all.