Influencer Analyzer

Scores influencers for partnership fit from real profile metrics and recent posts: audience quality signals, engagement authenticity, content stats, brand safety (caption keyword scan) and an overall partnership potential. Optional AI content assessment with your own OpenAI or Anthropic key.

---

Important: where the data comes from

This actor does not scrape Instagram, TikTok, YouTube, LinkedIn or X itself. It analyzes data you give it:

• Per influencer in the input: followers, following, postsCount and posts (recent posts with caption, likes, comments, and optionally url, timestamp, type), or
• A discovery dataset (discoveryDatasetId) whose items already carry real metrics, for example from the influencer-discovery actor run with demo mode off.

If an influencer has no usable data, the actor writes an error row (status: "error", errorCode: "NO_PROFILE_DATA") explaining what to supply, and does not charge for it. It never fills in placeholder numbers. Items that came from a demo-mode discovery run are skipped the same way (errorCode: "SAMPLE_SOURCE_DATA").

Any score the supplied data cannot support is returned as null and listed in unavailableFields.

---

Quick start

Real analysis with supplied data

{
  "influencers": [
    {
      "platform": "instagram",
      "username": "some_creator",
      "followers": 12000,
      "following": 800,
      "postsCount": 300,
      "posts": [
        { "caption": "New drop #style link in bio", "likes": 400, "comments": 20 },
        { "caption": "Morning run", "likes": 350, "comments": 15 },
        { "caption": "Giveaway time #giveaway", "likes": 900, "comments": 120 }
      ]
    }
  ],
  "analysisModules": ["audienceQuality", "engagementAuthenticity", "contentStyle", "brandSafety"]
}

Preview the output format (demo mode)

Set "demoMode": true. Every row is labeled isSampleData: true with random numbers that do not describe the real account. Demo rows are not charged.

---

What each module needs

Module	Needs	Output
audienceQuality	followers and following	Heuristic score from follower/following ratio and post count
engagementAuthenticity	3 or more posts with likes/comments	Like-to-comment ratio, consistency, spike detection
contentStyle	posts	Content types, hashtag and caption stats (no style score)
brandSafety	post captions	Keyword scan with your flagKeywords added to the defaults
growthTrend, competitorOverlap, sponsorshipHistory	not implemented yet	Listed in unavailableFields

analysisDepth: "comprehensive" with aiProvider and a matching API key adds an AI-written content assessment based on the supplied posts.

---

Input parameters

Parameter	Type	Default	Description
influencers	array	-	Influencers to analyze, with optional real metrics and posts
discoveryDatasetId	string	-	Dataset ID from the influencer-discovery actor
analysisDepth	string	"standard"	quick, standard or comprehensive
postsToAnalyze	integer	12	Maximum supplied posts analyzed per influencer
analysisModules	array	4 core modules	Which modules to run
brandContext	object	-	Passed to the AI assessment
aiProvider	string	"none"	openai, anthropic or none (BYOK)
openaiApiKey / anthropicApiKey	string	-	Your key for the AI assessment
flagKeywords	array	-	Extra brand safety keywords
demoMode	boolean	false	Return labeled sample data instead of analyzing
webhookUrl	string	-	Receives results and errors when the run finishes

---

Pricing

Pay per event:

Event	When it is charged	Price
Influencer Analyzed	An influencer with at least 2 modules scored from real data	$0.05

Not charged: demo mode rows, error rows, and partial rows where fewer than 2 modules could be scored (these are marked charged: false). Apify's standard actor start and per-result platform events still apply.

---

Output fields

status (ok, error or sample), isSampleData, dataSource, charged, profile metrics, overallScore (average of the scores that could be computed), per-module scores and details, partnershipPotential (high, medium, low, not_recommended or insufficient_data), strengths, concerns, recommendations, suggestedCollabTypes, unavailableFields.

---

Built by John Rippy | Actor Arsenal