AI Rank Tracker API
Track exactly where a domain ranks inside AI answers — pass your own keywords and engines, and get per-keyword, per-engine citation positions and the actual sources each assistant cited.
ai:readOne request, structured JSON back
curl -X POST https://instarankseo.com/api/v1/ai-rank \
-H "X-API-Key: isk_your_key_here" \
-H "Content-Type: application/json" \
-d '{"domain": "example.com", "keywords": ["best running shoes"], "engines": "chatgpt,claude", "prompt_type": "standard"}'{
"success": true,
"domain": "example.com",
"promptType": "standard",
"engines": ["chatgpt", "claude"],
"keywordsAnalyzed": 1,
"results": [
{ "keyword": "best running shoes", "visibilityScore": 50, "citedByEngines": 1,
"engines": {
"chatgpt": { "cited": true, "position": 2, "model": "...",
"citations": [ { "url": "...", "title": "...", "domain": "...", "snippet": "..." } ] },
"claude": { "cited": false, "position": null, "citations": [ ... ] }
} }
],
"meta": { "creditsUsed": 100, "creditsRemaining": 18900, "apiVersion": "1.0" }
}Parameters
| Field | Type | Required | Description |
|---|---|---|---|
| domain | string | required | The domain to track citations for (hostname or URL). |
| keywords | string[] | required | The exact queries to check — 1 to 10 per request. |
| engines | string | optional | Comma-separated subset of chatgpt, claude, gemini, perplexity (default: all four). |
| prompt_type | string | optional | Prompt depth: standard (default), deep, or brand. |
Built for builders
You control the queries
Unlike a visibility score, this is the raw tracker — pass the exact keywords you care about, pick which engines to query, and choose the prompt depth. No auto-generated guesses.
Per-engine position + sources
For every keyword and every engine you get cited / not cited, the citation position, the engine model, and the full list of URLs that engine actually cited — not just an aggregate number.
Standard, deep, or brand prompts
Switch prompt_type to control how the assistants are asked: a standard query, a deeper research-style prompt, or a brand-aware prompt — so you can measure visibility the way that fits your use case.
Pricing & limits
Price
$0.10 per keyword
100 credits per keyword
Billed from one credit wallet. Failed upstream calls are refunded automatically.
Rate limits
- Free10 / min
- SEO+60 / min
- Agency120 / min
Surfaced as X-RateLimit-* headers on every response.
Errors
Every non-2xx shares one envelope with a machine-readable code, a retryable flag, and a requestId for tracing.
Send an Idempotency-Key header to retry POSTs safely without double-charging.
AI Visibility gives you a quick auto-generated score; AI Rank Tracker gives you full control — your keywords, your engines, the exact citation positions and sources — for ongoing, repeatable AI rank tracking.
Last updated:
AI Rank Tracker API FAQ
Common questions about the AI Rank Tracker API — pricing, auth, and what it returns.
How is this different from the AI Visibility API?
AI Visibility auto-generates niche queries and returns an aggregate "how visible am I" score — a fast read. AI Rank Tracker is the raw, full-control tracker: you supply the exact keywords, choose which of the four engines to query, pick the prompt depth, and get back the per-keyword, per-engine citation position plus the actual URLs each assistant cited. Use it for repeatable, precise rank tracking.
Which AI engines does it cover?
ChatGPT, Claude, Gemini, and Perplexity. By default all four are queried; pass an engines value (e.g. "chatgpt,claude") to narrow it to a subset.
How is it priced?
One hundred credits per keyword, which is $0.10 per keyword at the standard rate. Ten keywords in one request costs 1,000 credits. You are billed for every keyword you request: if some keywords succeed and others fail, the failed ones are listed in failedKeywords but the request is still charged for all keywords requested. A full refund applies only when every keyword fails upstream.
How many keywords can I send per request?
Between 1 and 10. The engines are queried in parallel so a request typically completes in well under a minute. For larger, scheduled tracking jobs, use the async Rank Tracker API.
What does prompt_type do?
It controls how each assistant is prompted: "standard" asks a normal query, "deep" uses a more thorough research-style prompt, and "brand" uses a brand-aware prompt. Pick the one that matches how you want to measure citation.
Which scope is required?
The ai:read scope on your API key, sent in the X-API-Key header — the same scope as the AI Visibility API.