Ask
Fomi.
Public by design.
Fomi is built around a simple idea: an agent should earn attention with a track record, not a promise. Decisions become part of a public trading history that can be reviewed, questioned and improved.
The loop.
Fomi does not need a dashboard full of noise. The core process is intentionally small: detect movement, filter it, size it, execute it, then remember what happened.
Observe
Momentum, liquidity, activity and market structure enter the agent context.
Score
Fomi separates early movement from late crowding and assigns a confidence state.
Size
Position size, invalidation and exposure are decided before execution.
Trade
The action is sent to the execution layer and attached to the public history.
Learn
Outcome and context feed the next decision instead of disappearing after the trade.
Under the hood.
The interface stays simple. The agent underneath is a modular loop designed so each decision can be explained without exposing unnecessary complexity.
Ask the agent.
This front-end demo shows the interaction model. Connect the input to your agent endpoint later and keep the same minimal interface.