Enterprise-grade FP&A platform with probabilistic forecasting and AI insights. v2.0 delivers the architectural foundation; v2.1 completes the precision and intelligence layer.
RunwayX v2.0 represents a complete architectural transformation from a forecasting tool to an enterprise platform. This release establishes the foundation for production deployment with security, governance, and extensibility built-in.
v2.1-accuracy-refactor.
| Component | Status | Description |
|---|---|---|
| 🎨 Enterprise UI | ✅ Ready | Modern Streamlit interface with dark mode, 8 tabs, guided onboarding |
| 📊 Data Infrastructure | ✅ Ready | Multi-format upload (CSV/Excel/Parquet), Pydantic validation, quality scoring |
| 📈 Visualization Engine | ✅ Ready | 10+ interactive Plotly charts with zoom, pan, export |
| 🔮 Scenario Planning | ✅ Ready | What-if analysis with 5 templates, side-by-side comparison |
| 📡 Monitoring & Alerts | 🟨 Foundation | Email/Slack/Teams integration framework (wiring in progress) |
| 🤝 Collaboration | 🟨 Foundation | PDF/PowerPoint export, shareable links (auth integration pending) |
| 🔐 Security & Auth | 🟨 Foundation | SSO, RBAC, encryption modules (full roll-out in v2.1) |
| 🏢 Multi-Tenancy | 🟨 Foundation | Company isolation architecture (production testing in v2.1) |
| Component | Status | Branch | Target |
|---|---|---|---|
| 📊 Monte Carlo Engine | 🔧 Verification | v2.1-accuracy-refactor |
Deterministic simulation with validated confidence intervals |
| 💰 AR Collections Model | 🔧 Calibration | v2.1-cox-calibration |
C-index >0.70 target |
| 📈 Revenue Recognition | 🔧 Testing | v2.1-asc606-validation |
Comprehensive test coverage |
| 🤖 AI Insights Engine | 🔧 Tuning | v2.1-ai-tuning |
Precision thresholds and false positive reduction |
| 🧠 Advanced ML Models | 🔧 Development | v2.1-ml-enhancement |
LightGBM, Prophet/ARIMA, segmentation |
RunwayX provides probabilistic answers (not point estimates) for critical finance questions:
- 📊 Runway Forecasting: Monte Carlo simulation aggregating AR, revenue, and burn uncertainty (v2.1: accuracy validation)
- 💰 AR Collections: Cox survival model predicting invoice-level payment probabilities (v2.1: production calibration)
- 📈 Revenue Recognition: ASC-606 compliant UBP with tiered pricing (v2.1: expanded pricing models)
- 🎯 Collections Policy: Ranked action queue optimized by expected value (v2.1: reinforcement learning)
- 🤖 AI Assistant: Natural language chat for financial insights (v2.1: enhanced context awareness)
- 🔮 Scenario Planning: Interactive what-if analysis with real-time comparison (production ready)
# Clone and setup
git clone https://github.com/satyamsjha/runwayx.git
cd runwayx
python3.11 -m venv .venv && source .venv/bin/activate
# Install (production features)
pip install -e .[all]
# Generate synthetic data
python samples/make_synth.py
# Launch v2.0 Enterprise UI
streamlit run app_v2.py --server.port 8503
# → http://localhost:8503
# Optional: Classic UI
# streamlit run app.py --server.port 8501
# Optional: API Server
# uvicorn runwayx.cli.api:app --reload
# → http://localhost:8000/docsNote: For production deployments, review V2_DEPLOYMENT_COMPLETE.md.
- ✨ 50+ new modules and 15,000+ lines of enterprise-grade code
- 🏗️ Production-ready infrastructure: Security scaffolding, monitoring framework, audit logging
- 🎨 Modern UI: Dark mode, animations, guided onboarding, mobile-responsive
- 📊 Advanced visualizations: Interactive Plotly charts with export capabilities
- 🤝 Collaboration features: PDF/PPT export, shareable links, comments
- 🔐 Security Modules: OAuth2 (Google/Microsoft), RBAC architecture, AES-256 encryption
- 📡 Monitoring Framework: Email/Slack/Teams alert system, webhook infrastructure
- 🏢 Multi-company Architecture: Tenant isolation design, subscription tier framework
- ✅ Data governance: Quality scoring, duplicate detection, lineage tracking
- 📈 Scenario planning: Interactive what-if analysis with templates
- 🧪 54 tests passing across all infrastructure modules
- 📚 Comprehensive docs: INSTALL.md, QUICKSTART.md, RUN_APP_V2.md
- 🐳 Docker ready: One-command deployment
- 🎯 Type-safe: 100% type hints with Pydantic v2 validation
Focus: Mathematical accuracy, model calibration, and AI enhancement
-
🎯 Monte Carlo Refinement
- Target: Deterministic simulation with reproducible results
- Target: Validated confidence intervals (bootstrap + Bayesian)
- Target: Performance <5s for 10,000 simulations
- Target: Comprehensive sensitivity analysis
-
💰 AR Model Production Release
- Target: Cox model C-index >0.70
- LightGBM ensemble for non-linear patterns
- Hierarchical models by segment/region
- Real-time drift detection and retraining
-
📈 Revenue Recognition Expansion
- Hybrid pricing models (flat + usage)
- Material rights and contract modifications
- Multi-currency support
- Advanced true-up calculations
-
🤖 AI Intelligence Layer
- Enhanced context-aware chat
- Predictive anomaly detection
- Automated scenario generation
- Reinforcement learning for collections policy
-
🔌 Data Connectors
- Stripe, Chargebee, Maxio integrations
- QuickBooks Online, Xero, NetSuite
- Real-time sync and incremental updates
- Distributed simulation engine
- Time-series database integration
- Advanced caching and materialization
- API rate limiting and quotas
runwayx/
├── ai/ # AI insights & chat (v2.1: tuning)
├── api/ # FastAPI endpoints
├── ar_models/ # Cox + LightGBM (v2.1: calibration)
├── cli/ # Typer CLI + FastAPI
├── collab/ # Export & sharing (foundation)
├── core/ # Config, logging, settings
├── data/ # Data management utilities
├── enterprise/ # SSO, RBAC, audit (foundation)
├── forecasting/ # Time-series forecasting (v2.1: refinement)
├── governance/ # Data quality & lineage (ready)
├── ingest/ # Data loading & upload (ready)
├── ml/ # Segmentation, anomaly (v2.1: tuning)
├── monitoring/ # Alerts, webhooks (foundation)
├── policy/ # Collections optimizer (v2.1: RL)
├── reports/ # Interactive charts (ready)
├── revrec/ # Revenue recognition (v2.1: validation)
├── samples/ # Demo data generators
├── services/ # Business logic services
├── sim/ # Monte Carlo + scenarios (v2.1: accuracy)
├── tests/ # Test suite (54 passing)
└── ui/ # Theme & onboarding (ready)
# Run full test suite
pytest
# With coverage
pytest --cov=runwayx --cov-report=html
# Run specific module tests
pytest tests/test_simulator.py -vCurrent Status:
- ✅ 54 tests passing (infrastructure & UI)
- 🔧 Mathematical validation tests in progress (v2.1 branch)
- 🎯 Target: 100+ tests with >90% coverage for v2.1
- Installation Guide - Setup and dependencies
- Quick Start - Get running in 5 minutes
- v2.0 Features - Complete feature list
- Running v2 App - v2.0 UI guide
- Deployment Guide - Production deployment
- Monitoring Summary - Alert setup
These areas are functional but undergoing refinement for v2.1:
- Monte Carlo: Outputs under revalidation; recommend validating against known scenarios
- AR Model: C-index currently ~0.65; calibration target >0.70 in v2.1
- RevRec: Hybrid pricing model testing in progress
- Security Roll-out: SSO/RBAC modules implemented; full integration tracked for v2.1
- Multi-tenancy: Architecture complete; production testing in v2.1
We welcome contributions! Priority areas for v2.1:
- Mathematical validation: Help verify simulation accuracy and model calibration
- Data connectors: Build integrations for billing systems (Stripe, Chargebee, etc.)
- ML enhancements: Improve forecasting models and anomaly detection
- UI/UX improvements: Enhanced visualizations and user experience
- Documentation: Tutorials, guides, and examples
See CONTRIBUTING.md for guidelines (if available).
MIT License - See LICENSE
Made with ❤️ for CFOs and FP&A teams who need probabilistic answers, not point estimates.
⭐ Star this repo if you find it useful! Your feedback helps us prioritize v2.1 features.
RunwayX is built on the principle of transparent uncertainty quantification. Rather than providing false precision with point estimates, we deliver probability distributions that reflect real-world variability. v2.0 establishes the enterprise foundation; v2.1 perfects the mathematical precision.