Back to Overview

Foresight

SEC Event Intelligence Platform

See what's coming

January 2026 Launch

"You see things; and you say, 'Why?' But I dream things that never were; and I say, 'Why not?'"

— George Bernard Shaw

"There is one thing stronger than all the armies in the world, and that is an idea whose time has come."

— Victor Hugo

Transform SEC Filings into Trading Intelligence

Speed

Events extracted within minutes of SEC filing

Structure

30 event types with magnitude and timing

Actionable

Risk alerts with quantified historical impact

6 Core Products

Event Feed API

Structured JSON feeds of SEC events by filing type or company. Real-time delivery via REST API.

  • By filing type (8-K, 10-K, 10-Q)
  • By company (portfolio monitoring)
  • Tiers: Delayed → Real-time
✨ DEC 3 DEMO READY

SEC Oracle

Ask questions in plain English. Our AI understands your intent, queries across all your data sources, and delivers insights in seconds.

The Innovation: AI generates structured intent (not unreliable SQL) → we query multiple data sources → AI synthesizes actionable insights

  • Natural language: Ask complex questions like you're talking to an analyst
  • 19 data sources: Events, stock prices, financials, insiders, news, transcripts
  • Beats Fintool: We own the data + AI synthesis = deeper insights
  • Demo example: "Show defaults + stock moves" revealed 75% had warning signs
FREE TIER

Pattern-Based Signals (V13)

Rule-based event pattern matching for stock predictions. Simple IF-THEN rules that Sales can give away to prospects. Lead generation tool for V9 upsell.

Example Pattern: Management action during distress → +94% recovery (Contrarian Rebound)

  • 4+ validated patterns with documented historical returns (+94% to +47,633%)
  • No ML required: Fast, interpretable IF-THEN rules
  • 55-60% expected accuracy (vs V9's professional 0.1518 correlation)
  • Patterns: Contrarian Rebound, Red Flag Recovery, Sentiment Divergence
  • Free to give away - no training costs, perfect for lead generation
🔬 IN DEVELOPMENT

Event Prediction (V11)

Major upgrade over V8: Scaled to 10-20 years of historical data (300K-800K samples vs 45K) for better rare event detection and economic cycle coverage. Includes 2008 financial crisis, COVID, and full bull/bear market cycles.

V8 Baseline (Current): 0.25 correlation, F1 0.64 on 45K samples (3 years). V11 targets 0.28-0.32 correlation (+12-28%) with 10x more training data.
  • Historical Scaling: 300K-800K samples (2005-2024) vs V8's 45K (2022-2024)
  • Rare Event Coverage: 10x more bankruptcy/delisting examples → better F1 (0.64 → 0.75-0.80 target)
  • Economic Cycle Diversity: Includes 2008 crisis + COVID + bull markets for robust predictions
  • Turnaround Improvements: 0.31 → 0.40+ correlation target (needs 12-month complete cycles)
  • Status: Planning phase - data generation & training scheduled for Dec 2025
✅ PRODUCTION READY

Company Intelligence & Stock Recommendations (V9)

XGBoost meta-learning system delivers professional-grade stock recommendations (0.1518 correlation) and 5-dimensional company intelligence scores for 2,117 companies.

Two Products: Stock recommendations (STRONG BUY/BUY/HOLD/SELL) + Company intelligence scores (Event, Fundamental, Regime, Return, Confidence)
  • 0.1518 correlation (professional quant fund level, +18.6% vs baseline)
  • 78 STRONG BUY recommendations (top 3.7% of 2,117-company universe)
  • 71 features: 48 event signals (your secret sauce), 18 fundamentals, 5 regime indicators
  • 5-dimensional scoring: Event, Fundamental, Regime, Return, Confidence (0-100 each)
  • Expected Sharpe ratio: 1.8-2.3 with monthly rebalancing (top-tier performance)
🔬 IN DEVELOPMENT

Alternative Data Intelligence (V14)

Extract proprietary alternative data from SEC filing tables that standard APIs don't provide. Access 50-100+ data tables per company including segment revenue, debt schedules, deferred revenue, CapEx breakdowns, and more.

Competitive Edge: Proprietary data not available in FMP, Bloomberg, or FactSet APIs. Forward-looking indicators that predict performance before it shows in standard financials.
  • Segment-level insights: Geographic/product revenue breakdown (detect growth before it shows in totals)
  • Debt maturity schedules: Refinancing risk timing (predict distress 6-12 months early)
  • Deferred revenue: SaaS bookings proxy (leading indicator 3-6 months ahead of revenue)
  • 50-80 new features: Customer concentration, A/R aging, inventory composition, CapEx breakdown, R&D pipeline
  • Expected impact: +10-20% model performance improvement (r=0.19-0.21 vs V9's 0.1745)
  • Status: POC scripts ready, validating extraction quality on 100 test companies
🧪 RESEARCH TOOL

Stock Movement Scanner (V15)

Find stocks with interesting price movements (drops, spikes, rebounds, breakouts) and analyze what preceded them. Perfect for testing prediction models and validating event detection.

Primary Use: Research and model testing tool. Potential equity product feature if successful.
  • 7,704 US stocks: 2015-2025 price/volume data (~15M daily records)
  • Movement patterns: Sudden drops/spikes, volume spikes, rebounds, breakouts, gaps
  • Fast queries: ~3 seconds per year of data (real-time), <100ms with precompute
  • Test prediction models: Validate against real historical cases (e.g., COVID crash, FDA approvals)
  • Event validation: Did SEC events predict these movements? Statistical correlation analysis
  • Status: Planning complete, implementation starting (research MVP in 1-2 days)
💡 PLANNING PHASE

IPO Performance Modeling (V13)

Predict IPO performance despite the ultimate cold-start problem: no trading history, limited financial data. Client-requested research project.

Client Request: Getting asked about IPO modeling capability. Early-stage feasibility study in progress.
  • The challenge: Pre-IPO companies have 0-3 years of data vs decades for mature stocks
  • S-1 filing mining: Extract events from prospectuses (business history, risk factors, MD&A)
  • Cross-sectional signals: Pricing dynamics, underwriter quality, VC backing, market timing
  • Prediction targets: Day 1 pop, quality scores, risk stratification
  • Success target: R² > 0.15 on Day 1 pop (competitive with academic literature)
  • Status: Planning phase, awaiting data audit and client scoping (feasibility: 2 weeks, MVP: 6-8 weeks)

Product Evolution

Jan
2026

Initial Launch

Rule-based signals + SEC Oracle

  • • Event Feed API
  • • Risk Alerts
  • • SEC Oracle
Q1
2026
✓ READY

Event Prediction (V11) 🔬

Major upgrade: 10-20 years historical data for better rare event detection

  • 300K-800K samples: vs V8's 45K (10x more training data)
  • Economic cycles: Includes 2008 crisis + COVID + bull markets
  • Target: 0.28-0.32 correlation (+12-28% over V8's 0.25)
  • Status: Planning phase, Dec 2025 data generation & training
Q2
2026

XGBoost Portfolio Manager (V9)

Two-stage meta-learning with regime awareness

  • • Combines V11 event signals with fundamentals & regime data
  • • XGBoost learns non-linear feature interactions
  • • Insider trading integration for sentiment signals
  • • SHAP interpretability for transparency
Q3
2026

Alternative Data Intelligence (V14) 🔬

Proprietary SEC table extraction for competitive edge

  • 50-80 new features: Segment data, debt schedules, deferred revenue, CapEx
  • Forward-looking: Leading indicators 3-12 months ahead
  • Not in APIs: Data not available in FMP, Bloomberg, FactSet
  • Expected impact: +10-20% model improvement (r=0.19-0.21)

Foundation Data

11.9M
SEC Events Processed
267k
Filings Analyzed
30
Event Types Detected

Ready for Launch

Targeting hedge funds, asset managers, and research analysts. Beta program starts December 2025.