What it does
Explores an agent runtime where Jev selects from bounded, pageable actions while local code manages context and execution.
Benchmarks & research
Only you can see your notes.
Screenshot unavailable. Open the experiment ↗
Explores an agent runtime where Jev selects from bounded, pageable actions while local code manages context and execution.
For your own assistant, borrow the idea of choosing from a limited menu of actions. A developer can try the basic version without an AI model first. Have the program check each choice before running it. Keep a person in charge of reviewing changes to files or actions that affect the computer.