API

The Stock Sentiment API

Per-stock sentiment as data, not a chart. SentiSense reads the conversation around each covered US stock across news, Reddit, X, Substack, and finance YouTube, scores the tone of every mention with a model trained on financial text, and serves the result as a time series you can pull with one HTTP call.

The sentiment endpoints work on the free tier: 1,000 requests a month, no credit card. Generate a key and this works:

curl -H "X-SentiSense-API-Key: ss_live_YOUR_KEY" \
  "https://app.sentisense.ai/api/v2/metrics/entity/NVDA/metric/sentiment"

Also in the same API: Market Mood (a 0 to 100 market-wide composite), holdings-weighted ETF sentiment, analyst ratings, insider trades, and institutional 13F flows. Quotes are live market data; sentiment and news analytics are recomputed on a regular cadence through the day.

Want to see it before you build? The free market sentiment page renders the same data as a live sector heatmap, a most bullish and most bearish leaderboard, and weekly movers.

import requests

r = requests.get(
    "https://app.sentisense.ai/api/v2/metrics"
    "/entity/NVDA/metric/sentiment",
    headers={"X-SentiSense-API-Key": "ss_live_YOUR_KEY"},
    params={"maxDataPoints": 30},
)
for point in r.json():
    value = point["metricValue"]["value"]["value"]
    print(point["timestamp"], round(value, 3))

const res = await fetch(
  "https://app.sentisense.ai/api/v2/metrics" +
    "/entity/NVDA/metric/sentiment?maxDataPoints=30",
  { headers: { "X-SentiSense-API-Key": "ss_live_YOUR_KEY" } }
);
const series = await res.json();
// series[i].metricValue.value.value -> polarity

Response shape (trimmed)

[
  {
    "timestamp": 1782360000000,
    "metricType": "SENTIMENT",
    "metricValue": {
      "value": { "value": 0.1256 },
      "stats": { "mean": 0.1256, "min": 0.1256, "max": 0.1256, "count": 1 }
    }
  }
]

Prefer a typed client? Official SDKs: pip install sentisense and npm install sentisense. Both are thin wrappers over the same REST endpoints.

Methodology

Sentiment from the text itself, not inferred from price.

Most “sentiment” numbers are price-derived: momentum and volatility dressed up as mood. SentiSense works from the other direction. The pipeline ingests financial news, Reddit, X, Substack, and finance YouTube, resolves which company each mention actually refers to, and scores the tone of every individual mention with a model trained on financial text. Those mention-level scores roll up into the per-stock series the API serves, and the distribution endpoints let you split any metric back out by source.

That distinction matters if you are building on it: text-based sentiment can move before price does, which is exactly the gap a price-derived gauge cannot see. The full pipeline is documented on the methodology page, and if you want the background reading, start with how to track market sentiment.