A natural-language interface for stock screening — converts user queries to screener API calls and refines filters through conversation.
In 2024 plenty of entries let the model answer numerical questions directly and fell apart under scrutiny. This one confined the model to translation, turning a question into a screener API call and letting the API produce the figures, which structurally narrows where a hallucination can enter. The conversational refinement matters too, since screening is never one question: investors arrive vague and tighten filters as results appear.
Summaries are written by this site. Project pages include demo videos, screenshots and the team's own write-up (videos may autoplay).