AEO data, not a dashboard
For every customer site, every tracked question and every engine.
| Data |
What you get |
| Mentions |
Whether each engine's answer named the brand |
| Sentiment |
How each engine described the brand |
| Rankings |
Where the brand appeared, with the full answer text and the cited domains |
| Sources and citations |
The domains and sources the engines cite |
| Visibility |
How often the brand shows up across your tracked questions |
| Competitors |
Who else showed up, and how each was described |
| Question and series performance |
Results per question and per engine over a date range you pick |
| Negative mentions |
Answers that described the brand badly, pulled out |
Per customer site, with most data also rolled up across the whole account.
Engines: ChatGPT, Google AI Overviews, Google AI Mode, Perplexity, Claude, Gemini and Meta AI. Every engine comes back in the same shape, refreshed daily on current plans.
White-label AI visibility data, with your name on it
You already sell to a roster: campaigns, property managers, local businesses, agency clients. They have started asking what AI says about them. This is how you answer without building the pipeline.
- No Knowatoa name in the data. The API returns raw data. Your customers never log into anything of ours.
- Your IDs on every record. Add your own
external_id to each site in the app or over MCP. The API returns it on site and account records, so data maps straight to the customer in your system.
- Whole roster at once. Questions, rankings, sentiment, sources, visibility and competitors also come as account-wide calls, so you don't loop through sites.
White-label here means you rebrand raw data in your own product or reports. There's no themed dashboard to hand out.
An AI search monitoring API you poll on your schedule
- REST API. JSON or CSV. One API key per account.
- Poll when you want. You pull: call the API on your own schedule. The data is refreshed daily on current plans.
- MCP server. Signs in with OAuth. Your team can pull the same data from Claude, ChatGPT and other AI assistants.
Looking for one Big JSON after the run? The account-wide calls are the closest thing: one call per data type covers every site.
Read the API v2 documentation
Which engines, how deep
Not every customer needs the same depth. A close race with a big budget wants every engine. A small one just wants to track.
Data is refreshed daily by default. Partners and white-label platforms get the rest scoped on a call: which engines, how often data updates, and how many questions. That fits procurement better than a card trial. No credit card, and no security review triggered by a signup form.
Starting at $399
Built around how many questions you track, the update frequency, and which engines. We'll scope it with you.
Ask us for it
No card form here. If a credit card on a new vendor sets off weeks of security review at your company, start with a call. We scope the setup with you. Tell us what you want to build and for how many customers. We'll show you the data on a real one.
FAQ
Does Google offer API access to its generative AI data?
Not separately. Search Console counts AI Overviews and AI Mode inside its regular Web totals, and the Search Analytics API has no separate type for them. Its newer Generative AI report shows combined impressions only. Knowatoa's AI visibility API is different: the data comes from asking the AI engines directly, not from Google's reports.
Do I need to use your app?
Only for setup, and you keep access to the app. Customer sites, competitors and questions are set up once, in the app or through the MCP server. After that, you can work entirely through the API.
Do you push data, or do I pull it?
You pull it. Poll the API on your own schedule. Data is refreshed daily on current plans.
Can I resell the data under my own brand?
Yes, as data. There's no Knowatoa branding in what the API returns. You build the customer-facing part.
How does setup work for partners?
We scope it with you on a call: which engines, how often data updates, and how many questions. That fits procurement better than a card trial.