The problem
Researching a stock means jumping between price charts, indicator dashboards, and a stream of headlines, then holding all of it in your head at once. It is slow, and it is easy to anchor on whichever source you read first.
How it works
Instead of one model trying to do everything, the work is split across agents that each own one kind of evidence.
Market data→
Indicators→
News→
Synthesis
Each agent covers one sourceone combined view
- Market data covers price action and fundamentals.
- Technical indicators summarize trend and momentum signals.
- News pulls recent coverage and what it implies for the company.
- Synthesis weighs the three and writes the final view, including where the sources disagree.
Why agents
Splitting the job keeps each prompt focused and makes the reasoning auditable: you can see what each agent found before the summary, instead of trusting a single opaque answer.