Market Intelligence for Autonomous Systems
An AI agent that can generate alpha pays for its own compute. OptionWhales provides the signal layer — validated, programmatic, continuously updated — so agents can act on institutional options flow without building the data infrastructure themselves.
The Opening Pressure signal survived a three-stage out-of-sample test across 188.5 million option orders and 1,250 trading sessions. It is not a hypothesis — it is a documented statistical finding available via API.
Signal Validation Summary — Out-of-Sample (416 sessions, ~20 months)
1.81
OOS Sharpe (best)
+107.9%
Cumulative L/S Return
Many → Few
Configs Tested
188.5M orders
Data Basis
Full methodology in Pan & Poteshman replication paper →
View all research →What the API provides
Validated Directional Signal
The Opening Pressure signal is derived from 188.5 million large option orders across 1,250 sessions. It uses proprietary filters to isolate institutional directional conviction from routine flow. It passed a three-stage out-of-sample validation framework — not curve-fitted, not back-tested to death. Surviving signals deliver OOS Sharpe ratios of 1.6–1.8.
REST API + WebSocket Streaming
GET /v1/flow/current, GET /v1/momentum/rankings, GET /v1/abnormal-trades/current — JSON, authenticated by API key, 60 req/min on Pro. WebSocket (wss://api.optionwhales.io/v1/ws/abnormal-trades) for event-driven agents that need tick-level alerts rather than polling.
OpenClaw Skill (Agent-Native)
Install the official OptionWhales skill in any OpenClaw-compatible agent runtime in one command: clawhub install option-flow. The skill surfaces real-time flow rankings, momentum, and abnormal trades as structured tool outputs — no parsing required.
Daily Strategy Portfolio
Each session, the Opening Pressure pipeline materializes a ranked long/short portfolio based on proprietary institutional flow filters. Available at GET /strategy/today. Returns are tracked close-to-close and published the following day.
Historical Session Archive
154+ sessions of structured option flow data with per-ticker intent scores, clustering labels, directional confidence, abnormal trade flags, and GEX levels. Available programmatically via API for agent fine-tuning, backtest replay, or retrieval-augmented context.
Research-Grade Documentation
Two published research papers document the data provenance and signal validation methodology. An agent integrating this API can cite its reasoning: 'Signal based on 188.5M orders, OOS Sharpe 1.81, validated across 1,250 sessions from 2021–2026.'
The long-horizon thesis
A self-sustaining AI agent needs a signal that earns more than it costs
As AI agents become more autonomous — executing long-horizon tasks, managing resources, making real-world decisions — capital becomes a constraint. An agent that can generate consistent risk-adjusted returns from validated market signals can fund its own inference costs, storage, and compute without external subsidy.
The math is straightforward: a portfolio generating Sharpe 1.6–1.8 on a $10K base at 30% annualized gross returns produces ~$3K/year. At current LLM API pricing, that covers millions of tokens of context per day. The signal cost (OptionWhales Pro) is $19/month.
This is not a hypothetical future. The signal exists today. The API is live. The research validation is published and citable. An agent integrating this as a tool in its action space gains access to a statistically grounded market edge — not a rule of thumb, not a heuristic, but a documented, reproducible finding derived from 188.5 million real orders.
The vision: AI systems that are financially self-sustaining because their decision-making is grounded in validated quantitative signals rather than synthetic reasoning alone.
Integration guide
# Free tier: 10 req/min, top 3 tickers
# Pro: 60 req/min, all endpoints, WebSocket
GET https://optionwhales.io/registercurl https://api.optionwhales.io/strategy/today \
-H "X-API-Key: YOUR_KEY"
# Returns: { "longs": [...], "shorts": [...],
# "session": "2026-04-16", "signal_name": "opening_pressure" }import websocket, json
ws = websocket.create_connection(
"wss://api.optionwhales.io/v1/ws/abnormal-trades",
header=["X-API-Key: YOUR_KEY"]
)
while True:
trade = json.loads(ws.recv())
# trade: { ticker, strike, expiry, premium, direction, signal_score }
agent.process_signal(trade)const res = await fetch(
'https://api.optionwhales.io/v1/momentum/rankings',
{ headers: { 'X-API-Key': process.env.OW_API_KEY } }
);
const { rankings } = await res.json();
// rankings[0]: { ticker, score, direction, cluster, conviction }API tiers
| Feature | Free | Pro ($19/mo) |
|---|---|---|
| Flow rankings | Top 3 tickers | All tickers |
| Abnormal trades | Top 5 per session | Full feed |
| Strategy portfolio (/strategy/today) | — | ✓ |
| Historical sessions (/strategy/session/*) | — | ✓ |
| WebSocket streaming | — | ✓ |
| Momentum rankings | Current only | Full + history |
| Rate limit | 10 req/min / 200/day | 60 req/min / 5,000/day |
| OpenClaw skill | Read-only preview | Full access |
Start building
Free tier available immediately — no credit card, API key in under a minute. Full strategy portfolio and WebSocket access on Pro.