Jev in production › Agent tool and action selection

jev-use (savka777)

Voice-driven macOS computer use where Jev picks the next operation and target from the Accessibility tree, with no screenshots, and the app executes and re-reads the screen. About 0.3–1.5 s per step (author).

Open on GitHub ↗

about 0.3–1.5 s per stepmeasured, as published by the source
Use
Agent tool and action selection
Industry
Productivity
Form
Open-source tool
Stage
Beta
Plugs into
macOS
Listed
2026-09-20
Found via
github
Repository
savka777/jev-use
Stars
122
Forks
19
Last push
2026-09-21
Language
Swift
License
MIT

The README opens with

Voice and typed computer use for macOS. You say what you want. Jev picks the next on-screen action. macOS performs it. No screenshots: the app reads the screen through the Accessibility tree.

In setup: save your TypeSafe API key (stored in the Keychain), then allow Accessibility, microphone and speech.

Hold Control–Option–Space, speak, release. Change this combination in Settings → Choose your voice shortcut → Change shortcut; press the key and any modifiers you want. Your choice is saved across launches. Escape cancels recording, and a conflicting shortcut leaves the previous choice in place. macOS-reserved combinations and modifier-only shortcuts are not supported. Escape cancels. Or type a command in the widget, or from a shell:...

Badge

For the project's own README, linking back here:

Listed in Jev in production

Also used for agent tool and action selection