Skip to main content

​
Overview

The --extract sentiment flag returns structured JSON output containing overall sentiment, key claims, and notable accounts discussing a topic. This is useful for understanding the tone and substance of conversations on X.

​
Use Case

Sentiment extraction is ideal for:
  • Tracking crypto and financial market sentiment
  • Understanding public opinion on products or companies
  • Identifying key narratives and their frequency
  • Mapping influential voices and their stances

​
Example: Bitcoin Sentiment Analysis

Get a structured sentiment snapshot of Bitcoin discussions on X: 
/grok-x "Bitcoin sentiment on X right now" --extract sentiment

​
Output

Grok-X returns a structured JSON response: 
{
  "overall_sentiment": "cautiously bullish",
  "sentiment_score": 0.41,
  "key_claims": [
    {
      "claim": "The ETF inflow numbers suggest institutional accumulation is quietly accelerating despite retail uncertainty.",
      "frequency": "high",
      "representative_handle": "woonomic"
    },
    {
      "claim": "The halving is already priced in. Anyone expecting a 2020-style run is going to be disappointed.",
      "frequency": "moderate",
      "representative_handle": "PeterLBrandt"
    },
    {
      "claim": "On-chain data shows long-term holders are not selling. Diamond hands thesis remains intact.",
      "frequency": "high",
      "representative_handle": "glassnodealerts"
    },
    {
      "claim": "Macro headwinds from Fed policy are suppressing the ceiling. $120K is the new realistic target, not $250K.",
      "frequency": "moderate",
      "representative_handle": "RaoulGMI"
    },
    {
      "claim": "Altcoins are bleeding against BTC. Capital is rotating into Bitcoin dominance as a flight to safety.",
      "frequency": "moderate",
      "representative_handle": "CryptoCobain"
    }
  ],
  "notable_accounts": [
    { "handle": "woonomic", "stance": "bullish", "influence_tier": "high" },
    { "handle": "PeterLBrandt", "stance": "skeptical", "influence_tier": "high" },
    { "handle": "saylor", "stance": "strongly bullish", "influence_tier": "very high" },
    { "handle": "PrestonPysh", "stance": "bullish", "influence_tier": "medium" }
  ]
}

​
Output Structure

The sentiment extraction output contains three main sections:
​
overall_sentiment
string
A qualitative assessment of the dominant sentiment (e.g., “cautiously bullish”, “bearish”, “mixed”, “neutral”)
​
sentiment_score
number
A quantitative score typically ranging from -1.0 (very bearish) to +1.0 (very bullish). Scores near 0 indicate neutral or mixed sentiment.
​
key_claims
array
An array of the most frequently repeated claims or narratives:
​
notable_accounts
array
Key accounts driving the conversation:

​
Interpreting the Results

In this Bitcoin example:
  • Overall sentiment is “cautiously bullish” with a score of 0.41 (moderately positive)
  • Dominant narratives include institutional accumulation, halving concerns, and on-chain holder behavior
  • Key voices range from strongly bullish (saylor) to skeptical (PeterLBrandt)
  • Frequency tags help identify which claims are most widespread vs. niche
The sentiment_score is calibrated relative to the typical tone of discussions on the topic. A 0.41 for Bitcoin represents moderate optimism, but the same score for a different topic might indicate different sentiment intensity.
  • --extract narrative — Extract broader narrative structures instead of sentiment-focused analysis
  • --extract claims — Focus specifically on claims with attribution, without sentiment scoring
  • --from / --to — Narrow the time window for sentiment analysis
  • --handles — Limit sentiment analysis to specific accounts

Build docs developers (and LLMs) love

Get started for freeTalk to us