API reference

AI Investment Committee API

Committee runs five lenses over a symbol: Growth, Technical/Momentum, Options/Volatility, Value, and Risk Manager. Each returns a stance and a 0 to 100 sub-score, composited into a LyraMind Score.

What it returns

The response includes condition, plain, confidence, and a lyramind_score object with growth, momentum, valuation, risk, options_opportunity, and composite. The agents array carries each lens with its stance, score, basis, and a one-line point.

It also returns persona-tailored approaches (compliance-safe, never a recommendation) and the trust envelope.

When to use it

Use Committee when you want a structured, multi-angle read on a name rather than a single narrative: a scorecard you can rank or compare. The composite is a weighted blend (momentum 30%, growth 25%, then valuation, risk, and options 15% each).

Results are cached 5 minutes per symbol. When history is insufficient it returns available:false with a note.

Notes and honesty

This is explicitly a technical, flow, and behavioral committee. The Value lens reads price extension, not valuation multiples, and Growth reads trend and participation, not earnings. Fundamental inputs are a stated planned addition, and every lens declares its basis. No buy or sell calls.

curl -s https://api.lyramindos.com/v1/committee -H 'Content-Type: application/json' -H 'x-api-key: YOUR_KEY' -d '{"symbol":"TSLA"}'
POST /v1/committee (also GET /v1/committee?symbol=) · MCP lyramind_committee

Questions

Does the committee use fundamentals?

No. It is a technical, flow, and behavioral committee. The Value lens reads price extension and Growth reads trend and participation. Each lens states its basis, and fundamental inputs are a planned addition.

Is the LyraMind Score a buy signal?

No. It is a composite of technical and behavioral sub-scores for context. The response carries a disclaimer and offers compliance-safe approaches, never a recommendation.

Are results deterministic?

Yes. The committee composes deterministic engine outputs and is offline-safe, so the same read produces the same scores.

Related

See the live demo →Request access