# StressGen — Full Documentation

> AI-powered financial stress testing platform for DFAST and EBA regulatory compliance.

StressGen generates stress scenarios rooted in six layers of context: regulatory guidance, live market data, market intelligence, statistical validation, historical precedent, and likelihood intelligence. Describe a hypothetical economic event in plain English, and the platform produces a complete, audit-ready set of risk factor shocks.

---

## 1. Getting Started

### Welcome to StressGen

StressGen generates stress testing scenarios rooted in six layers of context: regulatory guidance, live market data, market intelligence, statistical validation, historical precedent, and likelihood intelligence. Describe a hypothetical economic event in plain English, and the platform produces a complete, audit-ready set of risk factor shocks.

Every number in your scenario traces back to a source — a regulatory document, a market data feed, or a statistical model. This is not generic AI output. It is context-rooted stress testing designed for regulatory submission.

### Create Your Account

1. Go to the Login page and sign up with your work email.
2. Your organization administrator will assign you a role (see User Roles below).
3. Once approved, log in and you'll land on the Dashboard.

### Your First Scenario

Creating a scenario takes just a few steps:

1. Navigate to Scenarios and click "New Scenario."
2. Describe the economic event you want to model. For example: "The Federal Reserve raises interest rates by 100 basis points amid persistent inflation, triggering a broad equity sell-off."
3. Choose which data sources to include (economic indicators, news, market data, etc.).
4. Select the regulatory framework — DFAST, EBA, or PRA.
5. Click "Generate Scenario" and wait for the AI to produce results (typically 1-3 minutes).

**Tip:** The more specific your description, the better the results. Include details about the trigger event, expected severity, and any particular sectors or regions you want emphasized.

### Understanding Results

After generation completes, your scenario page shows:

- **Summary cards** — Total number of shocks, confidence score, and a breakdown by category.
- **Shock table** — Every affected risk factor with its category, current value, and projected change.
- **Visualizations** — Charts showing shock distribution by asset class, severity, and direction.
- **AI narrative** — A written explanation of why each shock was applied, grounded in the scenario description and supporting data.

### User Roles

Your organization administrator assigns one of three roles:

- **Viewer** — Read-only access to scenarios and results.
- **Analyst** — Create, edit, and generate scenarios. Run stress tests and analyze results.
- **Admin** — Full access including user management, scenario approval, and organization settings.

### Feedback & Bug Reports

StressGen is under active development and your feedback shapes the product. When reporting an issue, include the steps to reproduce it, what you expected to happen, what actually happened (including any error messages), and your browser and operating system. Send reports to support@stressgen.ai.

---

## 2. Dashboard

### Your Dashboard

The Dashboard is the first thing you see after logging in. It gives you a real-time overview of markets, your recent work, and key risk indicators — all in one place. Every widget pulls live data, so you always have an up-to-date picture of conditions relevant to your stress testing.

### Available Widgets

The dashboard includes 13 widgets:

| Widget | What It Shows |
|--------|---------------|
| Market Snapshot | Broad market overview — major indices, Treasury yields, credit spreads, inflation expectations, and a market regime indicator (Risk On / Risk Off / Mixed). |
| Commodities | Oil, gold, natural gas, copper — current prices with daily changes and trend arrows. |
| FX Rates | Major currency pairs (EUR/USD, GBP/USD, USD/JPY, etc.) with daily changes. |
| Sector Performance | S&P 500 sector returns — daily, weekly, and year-to-date — sorted by daily change. |
| Calendar | Upcoming regulatory deadlines and economic data releases with importance levels. |
| Earnings Calendar | Upcoming corporate earnings dates with analyst estimates. |
| Quick Stats | Your scenario counts by status: total, completed, draft, in review, computing, and failed. |
| Crypto | Bitcoin, Ethereum, and other major crypto assets with prices, daily changes, and market cap. |
| VIX Term Structure | VIX futures curve — a visual chart of expected volatility over time. |
| Activity Feed | Recent platform updates, feature releases, and regulatory alerts from data sources. |
| Recent Scenarios | Your five most recent scenarios with status badges and direct links. |
| Scenario Likelihood | On-demand ranking of your completed scenarios by how likely they are given current market conditions. |
| Firm-Wide Risk | Top risk factor exposures aggregated across all your portfolios, with optional strategy filtering. |

### Customizing Your Layout

Click the settings gear icon in the top-right corner of the dashboard to open the customization panel. From there you can:

- **Toggle** — Turn individual widgets on or off with the switch next to each name.
- **Reorder** — Drag widgets up or down using the grip handle to arrange them in your preferred order.

Your layout preferences are saved to your account and persist across sessions and devices.

### Tips

Market data refreshes approximately every 5 minutes. Economic data from sources like FRED may have a 1-2 day lag. Scenario stats update in near-real-time as you create and generate scenarios.

---

## 3. Scenarios

### What Are Stress Scenarios?

A stress scenario describes a hypothetical economic event and its projected impact on financial markets. Think of it as a "what if" analysis: what happens to stock markets, interest rates, currencies, and commodities if a specific economic shock occurs?

StressGen generates a complete table of risk factor shocks for each scenario, showing how every tracked financial variable would be affected. These results are designed to meet regulatory requirements for bank stress testing submissions.

Every shock in a StressGen scenario comes with full provenance: which regulatory document informed the parameters, which market data feeds provided the baseline, and which statistical model propagated secondary effects.

### Scenario Types

StressGen supports four scenario types:

- **Historical** — Based on real historical events (e.g., the 2008 financial crisis, the 2020 COVID crash). The AI uses the actual market impacts as a reference for calibrating shocks.
- **Regulatory** — Aligned with DFAST, EBA, or PRA supervisory frameworks. These follow specific regulatory guidelines for risk factor coverage and severity.
- **Hypothetical** — Forward-looking, market-intelligence-driven scenarios. You define the event, and the AI generates shocks based on current conditions and historical patterns.
- **Black Swan** — Extreme tail-risk events with low probability but severe impact. Think pandemic, sovereign default, or unexpected geopolitical escalation.

### Creating a Scenario

Every scenario starts with a description of the economic event you want to model. Write it in plain English — the AI will interpret your intent and determine the appropriate shocks.

Good scenario descriptions include:
- The triggering event (e.g., a central bank rate decision, geopolitical conflict, pandemic)
- Expected severity (mild correction vs. severe crisis)
- Time horizon and affected regions or sectors
- Any specific risk factors you want emphasized

**Example:** "A rapid unwinding of the yen carry trade following a surprise Bank of Japan rate hike to 1%, causing global equity market volatility and a flight to quality in US Treasuries."

### Scenario Workflow

Every scenario follows a lifecycle with clear statuses:

| Status | Meaning |
|--------|---------|
| Draft | Scenario created but not yet generated. You can still edit the description and data sources. |
| Computing | The AI is generating shocks. A real-time progress bar shows each step. |
| Completed | Generation finished successfully. Results are ready for review. |
| In Review | Submitted for review by a risk manager or approver. |
| Approved | Reviewed and approved. Ready for regulatory submission. |
| Submitted | Submitted to the regulatory body. |
| Changes Requested | Reviewer requested modifications. The scenario returns to editable state. |
| Failed | Generation encountered an error. You can retry or edit and regenerate. |

### Editing & Versioning

After generating a scenario, you can refine it:

- **Edit description** — Update the scenario narrative and regenerate shocks with the revised context.
- **Adjust data sources** — Toggle different data feeds on or off to see how they influence the results.
- **Version history** — Each regeneration creates a new version so you can track changes over time.

### Comparing Scenarios

Use the comparison feature to view multiple scenarios side by side. This is especially useful for:
- Comparing regulatory baseline vs. severely adverse scenarios
- Evaluating how different trigger events affect the same risk factors
- Reviewing how scenario shocks change when different data sources are enabled

### Interpreting Results

Each generated scenario includes:

- **Shock table** — The complete list of risk factor impacts, organized by asset class.
- **Confidence score** — An indicator of how well the generated shocks align with historical patterns and regulatory precedent.
- **Narrative summary** — A plain-language explanation of the scenario's economic logic.
- **Supporting data** — The real-world data points that informed the AI's decisions.
- **Likelihood ranking** — How probable this scenario is given current market conditions.

### Exporting & Sharing

Scenarios can be exported for use in downstream systems or regulatory submissions. Export options include structured data formats suitable for risk management platforms, as well as presentation-ready summaries for stakeholder review.

---

## 4. Portfolios

### What Are Portfolios?

Portfolios in StressGen represent your trading books or investment strategies. Rather than tracking individual holdings, StressGen tracks your portfolio's sensitivity to risk factors — how much your P&L would change if a given risk factor moves.

This sensitivity-based approach is the standard in institutional risk management. It lets you connect your portfolios directly to stress test scenarios to see projected P&L impacts without exposing individual position details.

### Portfolio Strategies

Each portfolio is assigned a strategy that describes its investment approach:

| Strategy | Description |
|----------|-------------|
| Long Equity | Long-only equity positions in stocks and indices. |
| Short Equity | Short-biased equity positions for hedging or directional bets. |
| Rates Trading | Interest rate products — bonds, swaps, futures. |
| Credit Trading | Corporate bonds and credit derivatives. |
| FX Trading | Currency pairs and FX derivatives. |
| Multi-Asset | Diversified strategies spanning multiple asset classes. |
| Macro | Global macro strategies driven by economic themes. |
| Options | Options and volatility-focused strategies. |
| Mixed | General-purpose or unclassified portfolios (default). |

### Creating a Portfolio

1. Navigate to the Portfolios page and click "New Portfolio."
2. Enter a name and optional description for your portfolio.
3. Select the strategy that best describes its investment approach.
4. Save the portfolio. You can then start adding daily snapshots.

### Snapshots & Exposures

A **snapshot** captures your portfolio's risk factor sensitivities on a specific date. You can have one snapshot per day per portfolio — matching the industry standard of end-of-day risk reporting.

Each snapshot contains **exposures** — one per risk factor your portfolio is sensitive to. Each exposure records:

- **Delta** — First-order sensitivity — how much your P&L changes per unit move in the risk factor.
- **Gamma** — Second-order sensitivity (convexity) — how delta itself changes as the risk factor moves.
- **Vega** — Volatility sensitivity — relevant for options and volatility-linked positions.

Exposure units depend on the risk factor type:

- **$/bp** — Dollars per basis point, for interest rates and credit spreads
- **$/1%** — Dollars per 1% move, for equities, FX, and macro indicators
- **$/pt** — Dollars per point, for volatility indices like VIX

When adding exposures, every risk factor is validated against the platform's registry. If a factor name doesn't match, you'll see a clear error listing the unrecognized entries so you can correct them.

### Analytics

The portfolio detail page includes analytics to help you understand how your risk profile evolves:

- **Time series** — Track a single risk factor's sensitivity over time.
- **Exposure matrix** — A pivoted view showing all your risk factor exposures across dates.
- **Aggregate exposure** — Your top risk factor exposures across all portfolios, with optional filtering by strategy.

### Audit Trail

All portfolio actions are logged. Every create, update, and delete event is recorded with the user, timestamp, and action details.

---

## 5. Stress Tests

### What Are Stress Tests?

A stress test calculates the projected P&L impact of a scenario's shocks on your portfolio's exposures. It answers the question: "If this scenario happened, how much would my portfolio gain or lose?"

Stress tests connect two things you've already built — scenarios (with their risk factor shocks) and portfolio snapshots (with their sensitivities). The platform applies each shock to the corresponding exposure to compute factor-level and total P&L impact.

### Running a Stress Test

1. Navigate to Stress Tests and click "New Stress Test."
2. Give it a name and optional description.
3. Select one or more scenario-portfolio pairs.
4. Choose a shock methodology (see below).
5. Click "Run" — results compute immediately.

### Shock Methodologies

When a scenario has multiple historical shocks for the same risk factor, the methodology determines which one is used:

| Methodology | How It Selects Shocks |
|-------------|----------------------|
| Worst Case | Picks the shock with the largest absolute magnitude for each factor. |
| Historical Min | Picks the most negative shock for each factor — the worst historical drawdown. |
| Historical Max | Picks the most positive shock for each factor — the largest historical rally. |
| Directional | Follows the net price trend direction. If trending up, uses max; if down, uses min. |
| Mean | Averages all available shocks for each factor — a moderate, central-tendency estimate. |

For regulatory submissions, **Worst Case** is the most conservative and commonly expected. For internal analysis, **Directional** or **Mean** can provide a more nuanced view.

### Understanding Results

After a stress test completes, you see:

- **Impact matrix** — A heatmap showing total P&L impact for each scenario-portfolio pair.
- **Category breakdown** — P&L impact grouped by risk factor category (equities, rates, FX, commodities, credit, volatility, macro, inflation).
- **Factor-level detail** — For each risk factor: the applied shock, your delta/gamma/vega exposure, and the resulting P&L components.
- **Impact drivers chart** — A ranked view of which risk factors contribute the most to total P&L impact.

### Comparing Results

StressGen supports two types of comparison:

- **Within a stress test** — Compare different scenario-portfolio pairs within the same run.
- **Across stress tests** — Compare items from different runs to see how results change.

When comparing two items, the platform decomposes the P&L difference into:
- **Shock contribution** — How much of the difference is due to different shock values.
- **Sensitivity contribution** — How much is due to different portfolio exposures.

---

## 6. Risk Factors

### What Are Risk Factors?

Risk factors are the financial variables that stress scenarios measure. They represent key indicators across the global economy — from stock market indices to interest rates, currencies, and commodity prices.

### Risk Factor Categories

StressGen tracks over 100 risk factors across eight categories:

| Category | Examples |
|----------|----------|
| Stock Indices | S&P 500, NASDAQ, FTSE 100, Nikkei 225, Euro Stoxx 50 |
| Interest Rates | US Treasury 2Y/5Y/10Y/30Y, Fed Funds Rate, ECB Rate, SOFR, TGCR, OBFR |
| Currencies | EUR/USD, GBP/USD, USD/JPY, USD Index |
| Commodities | WTI Crude Oil, Gold, Natural Gas, Copper |
| Credit Spreads | Investment Grade, High Yield, Emerging Market spreads |
| Volatility | VIX (equity volatility), MOVE Index (bond volatility) |
| Economic Indicators | GDP Growth, Unemployment Rate, Consumer Price Index |
| Policy Rates | Federal Funds Rate, ECB Deposit Rate, Bank of England Rate |

### How Shocks Work

A "shock" is the projected change to a risk factor under the stress scenario. Shocks are expressed differently depending on the risk factor type:

- **Percentage change** — Used for stock indices, currencies, and commodities (e.g., -15%)
- **Basis point change** — Used for interest rates and credit spreads (e.g., +150 bps)
- **Absolute change** — Used for indicators like unemployment rate (e.g., +2.5 percentage points)

### Primary & Secondary Shocks

StressGen distinguishes between two types of shocks:

- **Primary shocks** — The direct, first-order impacts of the stress event. These are proposed by AI economists through structured debate across multiple rounds.
- **Secondary shocks** — The knock-on effects that ripple through connected markets. Calculated using C-Vine copulas (Clayton, Gumbel, Student-t) based on historical correlations between risk factors.

### Severity Levels

| Severity | Description | Typical Equity Impact |
|----------|-------------|----------------------|
| Baseline | Expected economic path with no significant stress | -5% to +5% |
| Adverse | A meaningful economic downturn or market disruption | -15% to -25% |
| Severely Adverse | A deep recession or financial crisis comparable to 2008 | -30% to -50% |

### Risk Factor Registry

The Risk Factor Registry is the central reference for all tracked risk factors. You can browse the full list, filter by category, and see each factor's display name, category, and associated tickers. The registry supports aliases (e.g., "S&P 500", "SPX", and "SP500" all refer to the same factor).

---

## 7. Data Sources

### Overview

StressGen enriches AI-generated scenarios with real-world data from multiple sources. When you create a scenario, you choose which data feeds to include. Data sources are organized into five categories.

### Economic Data

- **FRED (Federal Reserve Economic Data)** — The primary source for US economic indicators: GDP growth, unemployment rate, inflation, Treasury yields, and hundreds of other series maintained by the Federal Reserve Bank of St. Louis.
- **FED (Federal Reserve)** — Direct Federal Reserve data including policy statements, meeting minutes context, and official economic projections.
- **ECB (European Central Bank)** — European economic data including Eurozone GDP, inflation, and ECB policy rate information.

### News & Search

- **Tavily (News)** — AI-optimized search engine that finds the most relevant financial news articles for your scenario topic.
- **Tavily (Web Research)** — Broader web research for analyst reports, commentary, and research papers beyond news headlines.

### Market Data

- **Yahoo Finance** — Real-time and historical prices for equities, ETFs, indices, commodities, and currency pairs.

### Social Sentiment

- **Reddit** — Monitors financial communities for retail investor sentiment, trending topics, and emerging market narratives.

### Prediction Markets

- **Kalshi** — US-regulated prediction market offering probabilities on economic events, Fed rate decisions, and market milestones.
- **Polymarket** — Decentralized prediction market covering a wide range of economic and geopolitical events.

### Configuring Data Sources

When creating or editing a scenario, toggle individual sources on or off:

- **Enabled by default:** FRED, Tavily, and Yahoo Finance
- **Optional sources:** Reddit sentiment, prediction markets, alternative news providers
- **Coming soon:** Bloomberg, Reuters, and SEC EDGAR integrations

For regulatory scenarios, keep FRED and at least one news source enabled. For exploratory scenarios, try adding prediction markets and social sentiment for a broader perspective.

---

## 8. AI Assistant

### Overview

The AI Assistant is your conversational companion for exploring scenarios, understanding risk factors, and getting instant answers about your stress testing results. It's available from any page in the application.

### Opening the Assistant

- Click the chat button in the bottom-right corner of any page
- Use the keyboard shortcut Cmd+K (Mac) or Ctrl+K (Windows)

The assistant opens in a sidebar panel so you can keep your current view visible while chatting.

### What to Ask

- **Scenario analysis** — "What are the largest shocks in my latest scenario?" or "Why did the AI assign a -25% shock to equities?"
- **Risk factor questions** — "Explain why volatility tends to increase when stock markets decline"
- **Comparisons** — "Compare the Treasury yield shocks across my last 3 scenarios"
- **Regulatory guidance** — "What does DFAST require for credit spread stress testing?"
- **General help** — "How do I create a new scenario?" or "What data sources are available?"

### Using @Mentions

Reference specific items using @mentions:

| What to Reference | How to Type It | Example |
|-------------------|----------------|---------|
| A scenario | @scenario:name | "Summarize @scenario:Rate Hike 2025" |
| A risk factor | @factor:ticker | "What drives @factor:SP500 in this scenario?" |
| A regulatory folder | @folder:name | "What does @folder:DFAST 2025 say about credit risk?" |

The @folder: mention type is available on paid tiers for targeted regulatory document queries.

### Automated Actions

The AI Assistant can perform actions on your behalf:
- Looking up current market data for specific risk factors
- Querying regulatory documents for relevant guidance
- Searching your scenario history for relevant comparisons
- Generating summary statistics across multiple scenarios

### Tips for Better Results

- **Be specific** — Mention the scenario name or risk factor for more targeted answers.
- **Use @mentions** — They give the AI structured context for more accurate responses.
- **Ask follow-ups** — The assistant remembers conversation context.
- **Use Shift+Enter** — For multi-line questions that benefit from structured formatting.

The AI Assistant works best when you're on a specific scenario page — it automatically picks up context about the scenario you're viewing.

---

## Pages

- [Home](https://stressgen.ai)
- [Pricing](https://stressgen.ai/pricing)
- [About](https://stressgen.ai/about)
- [Documentation](https://stressgen.ai/docs)
- [Getting Started](https://stressgen.ai/docs/getting-started)
- [Dashboard Guide](https://stressgen.ai/docs/dashboard)
- [Scenarios](https://stressgen.ai/docs/scenarios)
- [Portfolios](https://stressgen.ai/docs/portfolios)
- [Stress Tests](https://stressgen.ai/docs/stress-tests)
- [Risk Factors](https://stressgen.ai/docs/risk-factors)
- [Data Sources](https://stressgen.ai/docs/data-sources)
- [AI Assistant](https://stressgen.ai/docs/ai-assistant)
- [FAQ](https://stressgen.ai/faq)
- [Contact](https://stressgen.ai/contact)
- [Privacy Policy](https://stressgen.ai/privacy)
- [System Status](https://stressgen.ai/status)