Value a company

Conversational. Give a ticker or describe a program; the agent runs a risk-adjusted NPV, benchmarks assumptions, and builds a downloadable model. Open the Inputs, Valuation, Catalysts, and Research tabs to see the full analysis, then refine by asking follow-ups (e.g. "what TPP would justify 30K peak patients?"). Reviewing the assumptions the agent surfaces is what makes the valuation defensible.

Screen an investment thesis

Guided and gated, for a development-stage public biotech. This guided flow walks four steps:

1. Consensus model — the market-implied, research-grounded valuation. Review it, then confirm.
2. Catalyst calendar — upcoming events. Add/edit/remove any, then confirm.
3. Scenarios — upside/downside levers on each catalyst.
4. Trajectory — probability-weighted price paths across the catalyst tree.

Human review is required, by design. The trajectory tree will not compute until you've confirmed the consensus model and the catalyst calendar — the "Analyst-reviewed" meter tracks your progress. This isn't a formality: your sign-off on the assumptions is what makes the probability-weighted paths trustworthy.