Ai Search Visibility Tracker

Track AI answer visibility for your brand across 500 prompts, weekly, by country — including which sources get cited and where you're missing.

• ✅ Target present/absent per query
• ✅ Competitors detected
• ✅ Cited domains + URLs extracted

What This Actor Does

• Monitors brand mentions in AI answers — Checks Perplexity, Google AI Overview, and more to see if your brand appears when people ask AI search engines the questions that matter to your business.
• Extracts cited sources — Captures every URL and domain the AI engine cites, so you know which sites are getting the visibility you want.
• Detects competitors — See which competitors show up in the same AI answers, and how often.
• Delivers structured, analysis-ready data — Every result is a clean JSON row with classification, match type, sources, and metadata — ready for spreadsheets, dashboards, or API workflows.

Why Not Just Ask ChatGPT?

Manually asking an AI engine one question at a time doesn't scale. This Actor solves three problems:

1. Scale — Test hundreds of prompts across multiple engines and markets in a single run.
2. Repeatability — Schedule weekly or daily runs and track how your visibility changes over time.
3. Structured outputs — Get machine-readable rows with classification, competitor detection, and source extraction — not just a chat transcript.

Key Features

• Batch prompts — Run 1 to 500+ queries per run
• Query templates — Auto-generate prompts from seed keywords using preset packs (agency, SaaS, ecommerce, local)
• Country & language targeting — 30 countries and 19 languages supported via dropdown selects
• Competitor detection — Track which rivals appear alongside or instead of you
• Citation extraction — Full source URLs, domains, and positions
• Query groups — Label queries by group for agency-friendly reporting
• Scheduling-ready — Connect to Apify Schedules for automated recurring monitoring

Templates

Choose a template to auto-generate queries from your seed keywords, or use custom to provide your own.

Template	Best For	Generated Patterns
custom	Full control	No generation — uses your queries as-is
agency	Broad funnel monitoring	"best {k}", "top {k} tools", "{k} vs {competitor}", "alternatives to {competitor}", "how to {k}", "best {k} for small business", "best {k} in {country}"
saas	Competitive SaaS visibility	"best {k} software", "{brand} alternatives", "{competitor} alternatives", "{brand} vs {competitor}", "best {k} for startups", "best {k} for agencies", "pricing comparison for {k} tools"
ecommerce	Product/category prompts	"best {k}", "top {k} brands", "best {k} for your needs", "{k} reviews", "where to buy {k}"
local	City-specific services	"best {k} in {city}", "top-rated {k} near me in {city}", "{k} cost in {city}", "emergency {k} {city}", "{k} reviews {city}"

Placeholders

Placeholder	Resolves To
{k}	Each seed keyword
{brand}	Your targetName
{competitor}	Each competitor (patterns skipped if none provided)
{country}	The country input value
{city}	Each location (local template only)

Input

Quick Start Examples

1. SaaS Template with Query Groups

Use a template with grouped queries for organized reporting. When queryGroups is provided, it takes priority over queries and template generation.

{
  "template": "saas",
  "seedKeywords": ["YouTube video"],
  "queryGroups": [
    {
      "groupName": "Video",
      "queries": ["Best video making tools"]
    }
  ],
  "targetName": "YouTube",
  "competitors": ["canva"],
  "engines": ["perplexity"],
  "country": "AU",
  "language": "en",
  "includeAnswerText": true,
  "includeSources": true,
  "includeCompetitors": false,
  "deduplicateSources": true
}

2. SaaS Template with Multiple Query Groups

Scale up with multiple query groups across different topics.

{
  "template": "saas",
  "seedKeywords": ["YouTube video"],
  "queryGroups": [
    {
      "groupName": "Video",
      "queries": [
        "Best video making tools",
        "YouTube Transcript",
        "AI Search"
      ]
    }
  ],
  "targetName": "YouTube",
  "competitors": ["canva"],
  "engines": ["perplexity"],
  "country": "AU",
  "language": "en",
  "includeAnswerText": true,
  "includeSources": true,
  "includeCompetitors": false,
  "deduplicateSources": true
}

3. Custom — Manual Queries

Provide your own queries for full control.

{
  "queries": [
    "best youtube transcript tools",
    "best tools to analyze youtube comments"
  ],
  "targetName": "Gitoken",
  "competitors": ["TubeBuddy", "VidIQ"],
  "engines": ["perplexity"],
  "includeCompetitors": true
}

4. Local Template — City-Specific Services

Monitor AI visibility for local services across multiple cities.

{
  "template": "local",
  "seedKeywords": ["plumber", "emergency plumber"],
  "locations": ["Sydney", "Melbourne", "Brisbane"],
  "targetName": "PipeFixPro",
  "engines": ["perplexity"],
  "country": "AU"
}

Input Parameters

Field	Type	Required	Default	Description
template	string	No	"custom"	Preset prompt pack: custom, agency, saas, ecommerce, local
seedKeywords	array	When using template	[]	Keywords to expand into queries via template patterns
locations	array	For local template	[]	Cities/areas for local query generation
queries	array	When custom	—	Prompts/questions to test against AI engines
queryGroups	array	No	[]	Grouped queries for agency reporting: [{ groupName, queries[] }]
targetName	string	Yes	—	Brand, product, or topic to track
targetAliases	array	No	[]	Alternate names, spelling variants
competitors	array	No	[]	Competitor names/domains to detect
engines	array	No	["perplexity"]	AI engines to query
country	string	No	"US"	Country (30 countries supported)
language	string	No	"en"	Language (19 languages supported)
includeAnswerText	boolean	No	true	Include full AI answer text
includeSources	boolean	No	true	Include cited sources
includeCompetitors	boolean	No	false	Enable competitor detection
deduplicateSources	boolean	No	true	Remove duplicate source URLs
saveScreenshot	boolean	No	false	Save page screenshots to KVS
saveHtml	boolean	No	false	Save page HTML to KVS
maxConcurrency	integer	No	1	Max concurrent browser pages (1–5)
navigationTimeoutSecs	integer	No	60	Page load timeout in seconds
maxRetries	integer	No	2	Retries for failed pages
proxyConfiguration	object	No	—	Optional proxy settings
debugMode	boolean	No	false	Verbose logging + debug info in output

Query Resolution Precedence

The Actor resolves queries in this order:

1. queryGroups provided → uses grouped queries, ignores queries and template generation
2. queries provided → uses them as-is, template only affects metadata labeling
3. Neither → generates queries from template + seedKeywords (+ locations for local)

Output

Table View

JSON Output

Each result row contains structured data like this real output from a production run:

[
  {
    "query": "Best video making tools",
    "queryGroup": "Video",
    "template": "saas",
    "engine": "perplexity",
    "country": "AU",
    "language": "en",
    "classification": "target_present",
    "targetName": "YouTube",
    "targetPresent": true,
    "targetMatchType": "exact",
    "sourceCount": 2,
    "answerLength": 1625,
    "timestamp": "2026-03-02T05:15:45.347Z"
  }
]

Full Result Row Fields

Field	Type	Description
query	string	The prompt that was tested
engine	string	Engine used (e.g., perplexity)
country	string	Country code
language	string	Language code
timestamp	string	ISO 8601 timestamp
aiAnswerFound	boolean	Whether an AI answer block was detected
answerText	string	Full extracted AI answer text
answerLength	number	Character count of answer
targetName	string	Brand being tracked
targetPresent	boolean	Whether target was found
targetMatchType	string	How target was matched: exact, normalized, alias, domain
targetMentions	array	Specific strings that matched
competitorsChecked	array	Competitors that were checked
competitorsPresent	array	Competitors found in answer
classification	string	Result classification (see below)
sources	array	Cited source objects with domain, url, title, position
sourceDomains	array	Unique domains from sources
sourceCount	number	Number of sources found
template	string	Template used for this run
templateVersion	string	Template version (e.g., "1.0")
queryGroup	string	Group label if using queryGroups
notes	string	Warnings or error messages
debug	object	Debug info (when debugMode: true)

Classification Values

Value	Meaning
target_present	AI answer found, target brand detected
target_absent	AI answer found, target brand NOT detected
no_ai_answer_found	Page loaded but no AI answer block found
partial_extraction	Extraction completed with warnings
error	Request failed entirely

Run Summary

Each run saves a RUN_SUMMARY to the Key-Value Store with aggregated metrics:

• visibilityRate — Fraction of successful AI answers where the target was present
• topCompetitors — Competitor mention counts (sorted)
• topSourceDomains — Most frequently cited domains (top 20)
• template, templateVersion, queryGroupCount — Template metadata
• Result counts: present, absent, no AI answer, errors

Interpreting Results

• Visibility rate — targetPresentCount / (results - errors - noAiAnswer). A rate of 0.40 means your brand appears in 40% of AI answers where an answer was found.
• Top cited domains — The topSourceDomains in the run summary shows which sites AI engines reference most. If your domain isn't there, consider improving content authority.
• Missing prompts — Filter for classification = "target_absent" to find the queries where your brand is NOT being recommended. These are your content gaps.
• Query groups — When using queryGroups, filter by queryGroup to compare visibility across different topic categories (e.g., "Brand awareness" vs "Competitor comparison").

Pricing

This Actor uses pay-per-event pricing: 1 event per query-engine pair processed (event name: query-result).

Scenario	Events
10 queries × 1 engine	10
10 queries × 2 engines	20
SaaS template, 3 keywords, 2 competitors, 1 engine	~21

Cost tips:

• Start with 1 engine (Perplexity) to validate prompts before adding more
• Use fewer seed keywords during testing
• Set includeAnswerText: false if you only need presence/absence classification (reduces storage)

Limitations

• Engine availability may vary — AI search engines can change their page structure or add bot detection at any time. The Actor includes retry logic, but some requests may still fail.
• Experimental engines — Engines labeled "experimental" (e.g., Google AI) have lower extraction reliability due to inconsistent AI Overview availability and aggressive bot detection.
• No LLM-based extraction — Answers are extracted via DOM selectors, not by calling an LLM API. This keeps costs predictable but means extraction depends on page structure.

Engine Support

Engine	Key	Status	Notes
Perplexity	perplexity	Stable	Primary engine, reliable extraction
Google AI Overview	google_ai	Experimental	Inconsistent availability, aggressive bot detection

Proxy

The Actor does not bundle proxies. If you need proxy support:

• Provide your own proxy configuration via the proxyConfiguration input
• Residential proxies recommended for Google AI to reduce blocking