Settings

Control autonomy, self-correction, model selection and insight delivery.

Autonomy

Autonomous mode

When on, the orchestrator wakes daily, scans new literature & trials, sets its own goal and runs the full pipeline. Last scan: 10/11/2026, 3:46:43 PM

Daily scan hour (UTC)

Earliest hour of the day the autonomous scan may fire.

Automatic self-correction

Run a self-correction cycle after every N completed runs. Last: 10/11/2026, 12:37:08 PM

Reasoning engine (swappable)

Research agents engine

Used by all ten agents in research runs, debates, reflections and bias audits.

Research Assistant default engine

Default for new chat conversations (can be changed per conversation).

Proactive insight delivery

Notify on accepted insights

Push a notification whenever the debate resolves an insight as accepted.

Novelty alert threshold

Also notify for non-rejected insights at or above this novelty score.

Appearance & data

Dark mode

The default is the light clinical theme.

Export settings

Download the current configuration as JSON.