WhaleRider

What the system does

Users define their own systematic rules and research hypotheses. WhaleRider turns those instructions into durable, inspectable definitions, validates them, runs historical simulations using market and financial data, and helps analyze simulated results. Open generated YAML in VS Code for direct control, or continue the user-directed workflow through an AI or the CLI.

Get your admin access key

Enter your workspace name and admin email, complete payment, and verify the code sent to that email. The resulting key is your workspace’s primary admin credential and is shown once.

1. Workspace details.Open Connect and provide your workspace name and admin email.
2. Complete payment.WhaleRider Early Access is $50 per month per workspace.
3. Verify your email.Enter the six-digit code sent to the admin email.
4. Save the key.Copy or download the key when it is revealed, then add the WhaleRider MCP URL to your preferred AI.

Add one MCP URL to the AI you already use

In ChatGPT, Claude, Gemini, or Grok, create a custom MCP connector named WhaleRider and use https://mcp.whalerider.org as the server URL. When WhaleRider opens its authorization page, enter your admin access key to approve the connection.

Once authorized, ask naturally: create a trade plan, refine it, run a simulation, and explore results.

Install and connect the WhaleRider CLI

irm https://cli.whalerider.org | iex

curl -s https://cli.whalerider.org/install-wr.sh | tr -d '
' | bash

Verify and create a profile

wr --version
wr profile set --name default --access-key <YOUR_ACCESS_KEY>
wr profile use --name default
wr profile list

Profiles are required for platform operations such as listing deployed artifacts or starting simulations. Local compilation is deliberately independent and does not require an access profile.

Install the WhaleRider DSL plugin

1. Open Extensions.Use the Extensions view inside VS Code.
2. Search for WhaleRider DSL.The extension provides the language tooling used by generated YAML documents.
3. Select Install.Open a WhaleRider YAML file to receive completion, hover documentation, diagnostics, and schema-aware validation while editing.

Four composable artifacts

Compilation validates each YAML document and creates a fixed .wr artifact. Deployment returns the identifiers used by the next composition layer.

Describe the complete decision process

Trade-plan risk controls

Define exactly what may be evaluated

A trade plan must supply MARKET_INDICES or TICKERS to bound its universe. Values inside one selector are alternatives, while supplied selectors are combined as restrictions. An ETF plan can use TICKERS: [CPER]; a ticker supplied alongside an index must also belong to that index at the evaluated time.

Supported source domains

Indicators can read market, company, event, macro, and position-history data. The table is the user-facing map of what may be measured and what may be filtered.

Complete financial measurement families

Financial domains contain many available fields, so they are grouped below by statement family. Expand a family to see every supported measurement name.

MEASUREMENT: SecuritiesTransacted * Price
FILTER: Side == Side.ACQUISITION

MEASUREMENT: Value
FILTER: EconomicIndicator == EconomicIndicator.FEDERAL_FUNDS

• Enum constants use the qualified form FieldName == EnumType.ENUM_MEMBER.
• Filters may combine conditions with AND, OR, and parentheses.
• DERIVED indicators reference prior indicator names and do not accept raw-record filters.
• Restrictive filters can yield missing periods; signals explicitly choose IGNORE, TRUE, or FALSE.
• ECONOMY uses Value and selects the series through EconomicIndicator.
• SECTOR_PE uses PE and selects the category through Sector.

Measurement and filter syntax

Indicators, transformations, signals, and state

Indicator fields

Each indicator has a unique name and domain. Depending on the source, it may include a measurement, interval, explicit ticker, raw-record filter, ordered steps, cross-sectional aggregation, relative time window, and evaluation interval.

Transformation library

Steps run in order, with each output becoming the next input. The language includes moving averages, oscillators, volatility, volume, pivots, channels, patterns, market structure, trend lines, ranking, bucketing, thresholds, and aggregation.

Signals

Signals are named Boolean expressions over indicators. They support arithmetic, comparisons, AND/OR, interval streaks, minimum and maximum streak bounds, and deliberate missing-data behavior.

Stateful criteria

Entry and exit logic can be a tree rather than a single condition. NEXT expresses sequence, sibling nodes express alternative branches, ABORT invalidates a route, FOR requires elapsed calendar time, TIMEOUT expires a route, and FREEZE_INDICATORS preserves setup or entry-time values for later decisions.

Control position size and aggregate exposure

With an ATR stop, the position-risk budget is divided by the entry-to-stop loss per share, including estimated commission. Without an ATR stop, WhaleRider sizes from capital limits and derives a stop price from the risk budget. Gaps, slippage, evaluation frequency, and simulated fill delay mean this is a sizing and control model, not a guaranteed-loss limit.

The simulation’s MAX_EQUITY_ALLOCATION is a separate gross-exposure cap for the strategy. All strategies share account equity and buying power. Allocations may exceed 100 to permit leveraged exposure or overlapping capacity, but they never create additional buying power.

Run one or more strategies in a shared account

Complete command surface

This reference is generated from the installed CLI using wr -h --all. Angle-bracket values are required placeholders unless shown in brackets.

Performance output

The performance command can group results daily, weekly, monthly, yearly, or across the full run. It returns trade counts, winners and losers, realized profit and return, extrema and averages, duration, position value, profit and loss, batting average, win/loss ratio, exposure, maximum drawdown, Sharpe, Sortino, profit factor, and average P&L per trade.

Continue from a completed run into research

A completed simulation is not the end of the workflow. Use the AI or CLI to retrieve authoritative summary metrics, group performance by time, inspect trades with pagination, isolate a difficult month, identify best and worst trades, and build annual-return charts, monthly heatmaps, or weekly equity curves.