Inspiration
Your Mac, by voice. Made for people who can’t see the screen.
I built Zoya for blind and low-vision people, with one mission: to make computers and the internet easier to use independently.
Playing a podcast. Shopping for an everyday essential. Sharing a thought online. People should be able to do these things without an inaccessible interface getting in the way.
With Zoya, you say what you want to do. She operates your Mac, explains what’s happening aloud, and asks for your approval before consequential actions.
Zoya works alongside tools like VoiceOver, helping people complete whole tasks through conversationwith their voice and on their own terms.
What it does
Zoya is a voice agent for the Mac. Users can ask her to find information, play a podcast or YouTube video, open Spotify, compose and publish a post, and navigate shopping and hotel-booking workflows.
She takes action and provides spoken feedback, so users can follow progress without reading the screen.
Control is central to the experience. Before consequential actions such as publishing or placing an order, Zoya asks for spoken confirmation. Users can interrupt with “Zoya, stop,” and she hands control back when passwords or one-time codes are required.
How we built it
We built Zoya in Python, with the Strands Agents SDK coordinating agents and tools for browser interaction, computer control, and task execution.
Speech recognition runs locally on the Mac using Whisper. Transcribed requests reach the agent, which uses tools to act. Amazon Polly
Built With
- amazon-web-services
- python
- strands
Log in or sign up for Devpost to join the conversation.