Perception Layer
The Web Speech API captures your voice and converts it to text in real time. A lightweight intent resolver routes commands like "Analyze XAU/JPY" or "Show me gold in euros" to the correct XAU cross-pair adapter (OANDA and Yahoo Finance bullion feeds).
Market Context Layer
- Multi-timeframe candle ingestion (1H and 15M).
- Structural mapping: PDH/PDL, equilibrium, premium/discount arrays.
- News & economic calendar awareness with high-impact filtering.
- Session and killzone awareness (London / New York GMT windows).
Reasoning Layer
A large language model with a 25-year institutional trader persona reasons over the prepared context using ICT and SMC playbooks — bias, sweep, displacement, OTE, OB, FVG, BOS/CHoCH — and outputs a structured plan with entry, stop, three targets, and an invalidation level.
Output Layer
The plan is rendered on dual lightweight-charts (1H + 15M) with zones drawn for FVGs, OBs, and liquidity, while text-to-speech narrates each step in sync with word-level orb pulses for a Jarvis-class voice experience.
Last updated · July 2026