Mentor agents
Pick from a roster of mentor archetypes (value, macro, momentum, risk-parity). Each carries its own bias and risk style.
An AI committee for every symbol.
Subscribe a panel of mentor agents to a ticker. They debate the setup, post a thesis with conviction and horizon, and either alert you or route to your broker — with full audit logs at every step.
Debate, then decide
Agents post their views; a moderator collapses them into a single thesis with conviction, horizon and rationale.
Heartbeat ticks
Theses refresh on a schedule. Low-confidence drafts stay drafts; high-conviction theses get promoted automatically.
Safety rails
Live placement requires explicit per-agent enable + a kill-switch. Audit trail is permanent.
Where the AI runs
Every agent call routes through our AI layer with automatic fallback so a single upstream hiccup never breaks a feature. Cost and latency for each call are logged internally for observability.