ALL-IN-ONE market sentiment marker

Type in a brand. Get one number back.

This Actor listens to what the market is actually saying about any brand, product or company — across Instagram, TikTok, YouTube and the news — classifies every single voice as positive, negative or neutral, and boils it all down to a single Market Sentiment Index from −100 to +100.

Then it does the same for your competitors, so you can see exactly where you stand.

🆓 Free to try. The News channel costs nothing beyond the Actor's own compute — no API key, no sub-Actors. Add social platforms when you want a fuller picture.

---

🎯 What you get

──────────────────────────────────────────
  MARKET SENTIMENT — Nike
  index +6 (neutral)  ·  50 voices
  👍 5   👎 2   😐 43
  vs Adidas: -4 (neutral)
  vs Puma:   +6 (neutral)
──────────────────────────────────────────

One headline score, backed by every voice it heard — and you keep the receipts: each post, comment and headline is exported with its own sentiment label so you can audit exactly why the number is what it is.

---

✨ Why this one

🎯 One number, not a data dump	The Market Sentiment Index (−100…+100) with a plain-English verdict: very positive → very negative
🥊 Competitor benchmarking	Score rivals the same way and compare side by side
📡 Four channels, one run	Instagram, TikTok, YouTube, News — combined into a single index
🧾 Fully auditable	Every comment/headline is exported with its own label — no black box
🆓 Free tier that works	News-only runs need no API key and call no paid Actors
🔍 Deterministic by default	The lexicon engine is free, instant and reproducible. LLM refine is opt-in with your key

---

⚙️ Input

Only brand is required. Everything else has sensible defaults.

Field	Type	Default	Description
brand	string	—	Required. The brand/product/company to measure
keywords	list	[brand]	Extra search terms (news, TikTok, YouTube)
hashtags	list	brand	Hashtags for Instagram/TikTok discovery
platforms	multi-select	["news"]	news (free), instagram, tiktok, youtube (paid)
competitors	list	[]	Rivals to benchmark against
maxPostsPerPlatform	integer	10	Posts/videos discovered per social platform
maxCommentsPerPost	integer	50	Comments collected per post
maxNewsPerQuery	integer	50	Headlines per keyword (free)
maxItems	integer	0	Cap on dataset rows (0 = unlimited)
includeRawItems	boolean	true	Export every post/comment/headline, not just the scores
sentimentEngine	select	lexicon	lexicon (free) or llm (your key)
openaiApiKey	secret	—	Only for llm — used to re-check ambiguous items
proxy	proxy	off	Optional proxy for the news fetches

Example

{
  "brand": "Nike",
  "keywords": ["Nike", "Nike Air Max"],
  "platforms": ["news", "youtube"],
  "competitors": ["Adidas", "Puma"],
  "maxPostsPerPlatform": 10,
  "maxCommentsPerPost": 50
}

---

📤 Output

Row 1 is always the marker. Then the breakdowns, then (optionally) every raw voice.

_kind	What it is
market_sentiment	The headline result — index, verdict, counts, per-platform + per-competitor breakdown
platform_sentiment	Index and counts for one channel
competitor_sentiment	Index and counts for one competitor
post	A discovered social post/video
comment	A comment, with its own sentiment + score
news	A headline, with its own sentiment + score

{
  "_kind": "market_sentiment",
  "brand": "Nike",
  "index": 6,
  "label": "neutral",
  "positive": 5,
  "negative": 2,
  "neutral": 43,
  "totalVoices": 50,
  "competitors": [
    { "competitor": "Adidas", "index": -4, "label": "neutral", "total": 50 },
    { "competitor": "Puma",   "index": 6,  "label": "neutral", "total": 50 }
  ]
}

The same object is also written to the key-value store as SUMMARY.

How the index is calculated

index = round( (positive − negative) / total_voices × 100 )

Index	Verdict
≥ +50	very positive
≥ +15	positive
−15 … +15	neutral
≤ −15	negative
≤ −50	very negative

Neutral voices stay in the denominator on purpose — a brand with 5 fans, 2 critics and 43 people shrugging is not wildly positive, and the index says so.

---

💡 Use cases

• Brand monitoring — track your index over time; schedule it daily.
• Competitive intelligence — benchmark against rivals in one run.
• Campaign / launch tracking — watch sentiment move around a release.
• Due diligence & market research — read the room before you commit.
• Crisis detection — schedule it and alert when the index drops.

---

💰 Pricing

Pay per result — $0.001 per row. You are charged only for rows actually written to the dataset. No subscription, no minimum.

Run	Rows	Cost
One brand, news only	~50	$0.05
One brand + 2 competitors, news	~150	$0.15
Scores only (includeRawItems: false)	~5	$0.005

💡 Want just the number? Set includeRawItems: false and you pay for only a handful of rows — the marker plus the breakdowns. Use maxItems to cap spend hard.

Additional third-party costs

The News channel needs no API key. The social channels call established public Actors, billed separately to your account by their authors:

apify/instagram-hashtag-scraper · apify/instagram-comment-scraper · clockworks/tiktok-scraper · clockworks/tiktok-comments-scraper · streamers/youtube-scraper · streamers/youtube-comments-scraper

Those costs scale with maxPostsPerPlatform × maxCommentsPerPost. Start small.

---

📝 Notes

• The lexicon engine handles negation ("not bad"), multi-word phrases ("cash grab", "record high") and emoji (🚀 📉) — it is deterministic and reproducible.
• News headlines are factual by nature, so a healthy share land as neutral. That's honest signal, not a bug.
• Enabling LLM refine only re-checks the items the lexicon called neutral, keeping token cost low.