Type:
In-house R&D
Year:
2026
Platform:
Desktop
Status:
Active R&D — not released
Capabilities:
Agentic AI, browser automation
Enquiries:
Active R&D — not released
Inside the product
The work, on screen
Starting a task
The opening prompt accepts a goal, a question, or a URL.
Live page
The browser view remains visible while it is on a live YouTube page.
Action trail
The panel lists the goal, the plan, each action taken, and the handoff when a captcha stops the run.
The challenge
Every browser has bolted AI onto the side of itself, and it shows. The tabs, the address bar and the back button are all still yours to operate, which leaves the assistant a passenger — helpful on occasion, but never actually driving.
What we built
Voya puts the agent in the driver's seat. One field takes an address or an intention, and from there the web gets worked in front of you — at a pace you can follow, on a screen you can watch, with the wheel yours again the second you want it. A browser you talk to rather than one you operate.
Safe by design
It never signs in and it never touches your credentials. At any wall — a login, a checkout, anything that ought to be a human decision — it stops, hands you the controls, waits, and picks up exactly where it left off. Autonomy where it helps, hands off where it counts.
