Features

Market Brain: every signal, one read.

Market Brain consumes every signal LyraMind computes (regime, participation, mean-reversion stretch, plus optional options, news, and flow context) and returns one coherent market read instead of a wall of indicators. It outputs meaning, not raw values.

Synthesis, not a dashboard

The engine derives component probabilities for trend, sideways drift, breakout, exhaustion, and reversal, then resolves them into a single market condition (trending, compression before expansion, accumulation, distribution, stretched, or range bound) with a confidence score. It attaches a plain-English narrative and a ranked list of risks: late-stage momentum, fading participation, statistical stretch, a recent liquidity sweep, elevated implied volatility, or nearby earnings.

What changed, and for whom

Market Brain runs a what-changed diff against recent sessions: regime shifts, volume spikes versus the 50-bar average, opening gaps, volatility jumps, and new range extremes. It also renders the same read through six investor lenses (long-term, income, trader, options learner, preserver, aggressive) as study angles, never instructions, so the interpretation matches who is reading it.

GET /market/brain · surfaced via GET/POST /v1/explain · MCP lyramind_explain

Questions

Where do the probabilities come from?

They are computed deterministically from the underlying engines (regime efficiency, participation quality, stretch signals). They express historical tendency and structure, not a forecast of price.

Can it give a buy or sell signal?

No. It describes condition, risk, and study angles for different investor types. It never recommends a trade or predicts a price.

What feeds the risk list?

Risks are generated only when the supporting signal is actually present: late-stage regime, weak participation, statistical stretch, liquidity sweep, high IV rank, or an upcoming earnings date.

Related

See the live demo →Request access