The Sentiment API quantifies social media activity and tone for digital assets, delivering sentiment scores and social volume metrics for use in models, monitoring and client-facing products.
Filtered for bot and spam activity, recorded point-in-time, and delivered by API or flat file.
Each record reports the level of social media discussion of an asset and the direction of its tone, together with the averages, ratios and rates of change required for modeling and monitoring.
Hourly, daily and long-term scores, each measured against the asset's own recent baseline. A reading above 50 indicates more positive discussion than the baseline period; a reading below 50, more negative.
Post volume and unique contributing accounts over a rolling 24-hour window, refreshed every minute, with bot and spam activity removed.
30-day averages, 7-day change and momentum, distinguishing a sustained shift in tone from short-term variation.
Social post volume per $1M of reported trading volume, identifying assets where discussion is elevated relative to trading activity.
{ "coin_uid": "bitcoin", "datetime": "2025-12-09T16:05:00.000Z", "hourly_sentiment": 53.97, "daily_sentiment": 42.28, "daily_sentiment_avg_30_days": 49.22, "daily_sentiment_momentum_7_days": 3.82, "long_term_sentiment": 32.04, "hype_to_activity_ratio": 1.86 }
The reliability of a backtest depends on its data. The history is preserved as originally recorded, retains assets that have since been delisted, and is available in bulk.
Every value is preserved as originally recorded, out-of-sample, so a backtest reflects the information available at the time.
Assets that have since been delisted remain in the history, reducing survivorship bias.
The complete history is available in bulk, at one-minute granularity from mid-2021 and at coarser intervals for earlier periods.
Sentiment and social volume serve as model inputs, with sufficient history to evaluate a strategy across multiple market cycles prior to deployment.
A material change in the volume or tone of discussion around a holding provides an early indicator for review alongside price and news.
Surface the assets attracting the most attention, and the direction of sentiment on each, within your own product and under your own brand.
Real-time news from filings, exchange notices, project blogs and the press, tagged by coin, company and event.
Vesting and unlock schedules, with every event dated, sized and attributed to its recipients.
Sector, subsector, ecosystem and attribute classification for 1,700+ digital assets.
Staking yields and network data, with independent risk ratings from AAA to D.
Quantitative funds use it as a model input; asset managers and research teams use it to monitor holdings and track narratives; exchanges and platforms use it to display trending assets and sentiment to their users.
Building and backtesting trading signals, adding social attention as a factor in existing models, monitoring portfolios for shifts in tone, tracking which narratives are gaining or losing attention, and powering trending lists and asset pages.
Social media posts referencing each asset are filtered for bot and spam activity and scored for tone. Each score then compares a recent window with a longer baseline: the last hour against the previous 24 hours, the last 24 hours against the previous 20 days, and the last 50 days against the previous 200. Full metric definitions are in the API documentation.
Yes. The history is point-in-time and out-of-sample, and it includes assets that have since been delisted, which reduces look-ahead and survivorship bias. It is delivered as flat files as well as through the API.
Yes. With the Taxonomy API, sentiment can be analyzed at sector level. With the News API, a change in tone can be read alongside the events behind it.
Yes. Support can be added for any tradeable asset on request.
Contact our team to discuss your requirements. A member of the team will respond.