Extracting signal from noisy sources through rules-based structuring. Allowing you to factor in sources that previously couldn't be added to your workflow.
Methodology
How the S-Score is built.
Nothing blended, nothing model-generated.
01 / ScoreFine-grain scoredMessages tagged with a security are scored from −1.0000 to 1.0000.
02 / WeightWeighted over 24 hoursScores are exponentially time-weight summed by security over a rolling 24-hour period.
03 / StandardizeAgainst a 20-day baselineA 20-day mean and standard deviation set the baseline, producing a score from −4.25 to +4.25.
04 / FlagSignal separated from noiseStatistical significance flags show which moves are meaningful.
message_score ∈ [−1.0000, 1.0000] → Σ exp-weighted (24h rolling) → z vs 20-day μ, σ → s_score ∈ [−4.25, +4.25]
Track record
A decade of history, proving alpha generation.
We publish the backtest and provide the history behind it during trials, so your team can run its own.
Daily X sentiment (S-Score), close-to-closePrice > $5 universe · 12/1/11 to 7/22/26
Historical, backtested results. Past performance is not indicative of future results. [DISCLAIMER: confirm wording with compliance]
One engine, three feeds
Same signal. Three ways in.
01S-Factor Feed
02Activity Feed
03AI Summary Feed
01 / 03
01 / S-Factor Feed
The core social signal
Core sentiment metrics with S-Score values, statistical significance flags and historical trend data, updated in real time.
S-Score values from −4.25 to +4.25Statistical significance flagsHistorical trend data and real-time updates
01 / S-FACTOR FEED
01 · S-Factor feedThe core social signal
02 / Activity Feed
The granular view
A time-slice view of social activity showing when raw data was ingested, for visibility into the timestamps of event catalysts.
Volume metrics and source attributionDaily sentiment scoresIngest timestamps for every catalyst
02 / ACTIVITY FEED
02 · Activity feedThe granular view
03 / AI Summary Feed
The signal, in plain English
A 24-hour English summary of the same discussion behind the S-Score. AI is added after to describe the signal. It never qualifies it.
Why a name is moving, without building a modelKey themes and narrative in the discussionScores stay rules-based, never model-generated
● 1-min
MSFTlast 24 hours
Specifications
Built to plug into your stack.
Factors15, including S-Score and S-Volume
Frequency1-minute, 15-minute, daily or custom
SourcesX and StockTwits, each licensed separately with its own history and coverage
HistoryStockTwits from mid-2009. X from December 2011 for U.S. equities
DeliveryJSON API, flat-file SFTP or custom, scoped to the asset classes and sources you select
U.S. equities5,500+
ETFs3,500+
Japan (JPX)3,500+
Pan-European1,000+
London (LSE)950+
Crypto850+
Australia (ASX)850+
India (NSE)800+
Private companies700+
Canada (TSX)450+
Futures100+
Forex pairs50
FAQ
Frequently Asked Questions
No. It is a rules-based NLP engine with 15+ qualifying factors and full transparency into source, processing and signal.
No. X and StockTwits are separate, independently selectable sentiment sources, not a single bundled feed. You choose which source (or both) to license based on your strategy, and each source carries its own coverage history and asset class availability.
Messages tagged with a security are fine-grain scored from −1.0000 to 1.0000, then exponentially time-weighted and summed by security over a rolling 24-hour period. A 20-day mean and standard deviation serve as the historical baseline for a standardized score from −4.25 to +4.25, flagged for statistical significance so you can distinguish signal from noise.
S-Factor is available at multiple signal frequencies — 1-minute, 15-minute, and daily — and we can configure custom frequencies so you can match the cadence to your strategy rather than being locked into one update speed.
U.S. equities (5,500+), ETFs (3,500+), futures (100+), forex (50 pairs), crypto (850+), private companies (700+), and international exchanges including TSX, NSE, LSE, ASX, JPX and Pan-European markets. Coverage is delivered by the asset class and source combination you need, rather than as one fixed universe.
Standardly via JSON API, scoped to the asset class and source combination you select, for direct integration into trading dashboards, GUIs and analytics platforms. Flat-file SFTP delivery is also available, and custom delivery can be accommodated.
Add social as your next signal.
Tell us your universe, sources and horizon. We will send a historical data cut and walk you through the methodology.