Repeat a consistent set of questions
Save brand aliases, tracked domains, competitors, engines, prompts and a country. Runs can repeat every 1–168 hours. Changing the prompt set or engines changes what you measure; retain those settings when comparing periods.
Separate mentions, citations and competitive share
Mentions match brand aliases in answer text; citations match registered domains in the source list. Compare per-engine mention rates and competitive share within the observed answer set. These are not estimates of total market share or traditional SEO rankings.
How to track your brand in ChatGPT
The example below runs the same prompt daily across ChatGPT and Gemini. Keep category discovery prompts separate from direct brand comparisons. Register the Korean and English brand names as aliases when both are used, and add the official sites you want to track to domains.
AI visibility metrics available through the API
Read GET /v1/monitors/:id to bring observed answer statistics into your reporting. Inspect the underlying answers and sources alongside aggregate metrics to understand a change.
- engineRates: compare mention and citation results by engine.
- topDomains, topPages: inspect cited domains and individual pages.
- shareOfVoice: calculate each brand’s mention rate from runs and its mentions count. Multiple brands can appear in one answer, so the rates need not sum to 100%.
Create a monitor
Send this input to POST /v1/monitors. Replace the illustrative brand and domain with your actual tracking target. You can also create, edit and pause monitors in the dashboard.
{
"name": "Earbud brand tracking",
"engines": [
"CHATGPT",
"GEMINI"
],
"prompts": [
"best wireless earbuds for commuting"
],
"aliases": [
"Acme Audio"
],
"domains": [
"example.com"
],
"competitors": [
{
"name": "Sony",
"aliases": [
"Sony"
]
}
],
"country": "US",
"intervalHours": 24
}Scheduled runs use ordinary task limits
Each run creates prompts × engines tasks. For example, 12 prompts across ChatGPT and Gemini create 24 tasks and consume 36 credits if all complete. Plan concurrency and available balance apply; a scheduled time does not guarantee every answer is ready immediately.
Inspect the evidence behind changes
When mention rates change, inspect the underlying answers and sources. Check whether prompt composition, execution failures or AI summary availability changed too. Domain citation tracking requires registered domains.