REST for your code, MCP for your AI assistants, alerts to your inbox — same scored options dataset, no drift.
GET /api/v1/uoa?symbol=NVDA&min_score=85
{ "data": [{
"symbol": "NVDA", "strike": 950, "type": "C",
"score": 96, "premium": 2350000, "tag": "sweep"
}] }
tools/call · uoa_lookup
{ "symbol": "NVDA", "min_score": 85 }
→ "NVDA $950C, sweep, 96/100 unusual, $2.35M premium"
WHEN unusual_score >= 90 AND symbol IN watchlist:mega-caps
THEN notify(sms, slack#flow-critical)
WITH payload(full_event)
Python, AI assistant, or just an alert when something unusual fires — same dataset, three doors in.
Four design choices that separate this from a basic data feed.
JSON over HTTPS, 100 req/min on Vega. Access chains, Greeks, 80+ derived metrics, and historical analytics.
Drop the MCP server into Cursor, Claude, or ChatGPT and let the assistant query options data in plain English.
Rules on any metric, fanned out to email, SMS, Slack, Discord, push, or your own webhook.
API, MCP, alerts, and workstation read from the same scored archive. What you query matches what you see.
A dozen vendors will sell you ticks. None ship with an MCP server and a 6-channel alert engine.
Everything we get asked about the API, MCP, and alert delivery.
Free to test the API and MCP. Gamma unlocks alert workflows; Vega unlocks production API and MCP rates.