TradeMCP News Docs

Asset sentiment

Sentiment time series per asset, and which assets are suddenly in the news.

All endpoints on this page need the sentiment:read scope.

Tracked assets

GET /v1/assets lists the symbols articles are tagged with.

Sentiment series

GET /v1/assets/{symbol}/sentiment?interval=1h|4h|1d&since&until

{"symbol": "BTC", "interval": "4h", "points": [{"ts": "…", "mentions": 41, "stories": 23,
  "positive": 15, "negative": 9, "neutral": 17, "sentiment_avg": 0.08, "sentiment_weighted": 0.11,
  "impact_max": 0.72}]}
  • mentions counts articles, stories counts unique stories: a story syndicated by 10 outlets is 10 mentions but 1 story.
  • sentiment_avg is the plain average score; sentiment_weighted weights each article by its impact.
  • Series are kept for a year, so they reach further back than the 30 days of searchable articles.

GET /v1/assets/trending?window_hours=24&limit=20 ranks symbols by volume_z: how far their mentions in the window are above their own 7-day baseline, as a z-score. It surfaces assets that are suddenly being talked about, whatever the direction.

Reading the scores

FieldRangeMeaning
score−1 to 1Bearish to bullish: probability positive minus probability negative
labelpositive, negative or neutral, whichever is most likely
confidence0 to 1How sure the model is of the label
impact0 to 1Score strength combined with source credibility, novelty and how central the asset is

On this page