API reference

Agent Marketplace API

Marketplace lists the roster of opinionated agent personas that run over the same engine: Investor, Compliance, Macro, and Catalyst. Each is a focused lens that composes the primitives for one job.

What it returns

The catalog returns an agents array where each entry carries id, name, tagline, and records_verdict, plus a note and powered_by.

records_verdict tells you which persona writes to the Trust Ledger: only the Investor persona runs a readiness check and records a verdict; the others are ledger-neutral reads.

When to use it

Call Marketplace to discover the available personas and their ids before invoking one via the run endpoint. It is an open catalog with no per-call metering, ideal for populating a picker or an agent's tool menu.

The four personas: Investor (readiness plus committee score), Compliance (auditable and grounded?), Macro (top-down factors and cohort), and Catalyst (risks, changes, events).

Notes and honesty

Every persona produces a study of considerations, verdicts, and factors, never a buy or sell recommendation. The composition is deterministic and offline-safe; an LLM-authored persona can plug in behind the same run contract later.

curl -s https://api.lyramindos.com/v1/marketplace
GET /v1/marketplace

Questions

Which persona records to the ledger?

Only the Investor persona, which runs a readiness check and records a verdict. Compliance, Macro, and Catalyst are ledger-neutral reads, as shown by the records_verdict flag.

Do I need a key for the catalog?

No. The catalog endpoint is open and not metered. The run endpoint that executes a persona is metered like the other primitives.

Do personas give recommendations?

No. Each produces a study of considerations and factors with the Trust Layer, never a buy or sell recommendation.

Related

See the live demo →Request access