The local tracker reports observed sample metrics. It does not produce a provider ranking or a universal AI visibility score.
Names are matched case-insensitively using text boundaries. The tool excludes URLs from text matching. Broad or ambiguous aliases can create false positives; review the original answer.
URLs come from response text and source URLs you supply. Exact hostnames and subdomains are accepted; lookalike domains are not. A URL match does not verify how a provider displayed a citation.
A response can mention multiple brands. Each brand contributes at most one presence per response. This is a share of your monitored name-presence counts, not market share, reach or recommendation rank.
Platform, model, region and observation date are supplied by you. Date charts show observed dates with saved records; missing dates are not treated as zero visibility.
All percentages describe the current filtered sample. Repeated prompts, changing models, personalised sessions and inconsistent collection conditions can affect comparisons. No confidence interval, ranking position, search volume or population-wide estimate is calculated.
Changing brand aliases or competitor settings recalculates old observations under the current rules. Preserve a separate report if you need to retain an earlier interpretation.
Example dashboard graphics are fictional and labelled. No live data is retrieved from AI providers by this package.
Organise your prompts, record the responses and see where your brand appears in your sample.
Explore the AI visibility tracker → Read the methodology
Arolag