███████╗ ███████╗ ██████╗ ██████╗ ╚══███╔╝ ██╔════╝ ██╔══██╗ ██╔═══██╗ ███╔╝ █████╗ ██████╔╝ ██║ ██║ ███╔╝ ██╔══╝ ██╔══██╗ ██║ ██║ ███████╗ ███████╗ ██║ ██║ ╚██████╔╝ ╚══════╝ ╚══════╝ ╚═╝ ╚═╝ ╚═════╝
▮
every market, evaluated continuously across independent lenses. your agent asks. zero answers in 200ms — verdict, sizing, confidence, every output traceable to the live signal beneath it. designed for AI to call directly.
“multi-lens evaluation. only setups where independent perspectives converge.”
“risk-adjusted to portfolio. stop attached. trailing managed. atomic execution.”
“continuously calibrated against live markets. signals that win get weighted up.”
“every output traceable. every outcome journaled. operating in production, not in slides.”
POST /api/v1/decide
designed for AI agents to call directly.
scout
→ { "coin": "BTC", "direction": "LONG", "horizon_h": 6 }
analyst
← { "verdict": "PASS", "confidence": 0.74, "agreeing_lenses": ["price","macro","flow"], "size_pct": 8.2, "stop_pct": 3.0, "max_hold_h": 6, "regime": "trending" }
shield
structured for tool-use. webhooks for execution. attribution baked into every output.
zero
your agent stops hallucinating about markets. it asks the engine running on them.
“the engine identified autonomously that 2-6h holds outperform longer ones. nobody encoded that. the loop discovered it. every subsequent decision inherits that knowledge.”
a continuously-evolving reasoning system. your agent does not consume static signals — it queries an engine that is sharper today than it was yesterday, and sharper tomorrow than today.
build what comes after the chatbot.
private beta — one key per developer. structured for tool-use. attribution on every output. running in production, not slides.