My Voice

A voice-first accessibility companion that helps vulnerable people understand information, remember what matters, communicate in their own words, and stay in control of what happens next.

Comment

Inspiration

A lot of technology for vulnerable people is designed around monitoring, supervising, or informing carers and family members.

My Voice starts from the opposite direction: build for the person themselves.

A person may receive a confusing bill, forget what they needed to follow up, struggle to explain a concern, or feel uncomfortable about a suspicious request. They should be able to ask for help without losing control of their own words or choices.

My Voice is designed to listen first.

What it does

My Voice is a simulated Alexa+ accessibility and advocacy companion designed for elderly, disabled, cognitively impaired, and otherwise vulnerable adults.

Its core interaction is:

What you said → What I understood → What do you want done?

The person's exact words are kept separate from the system's interpretation, and nothing is shared automatically.

The prototype provides six core workflows:

  • Help me understand — explains important information from a letter, bill, form, or message in plain language.
  • Remember this — keeps a follow-up or concern visible until the user decides it is finished.
  • Help me say this — preserves the user's exact words and shows a separate neutral interpretation before asking what they want done.
  • Something doesn't feel right — checks for warning signs such as requests for passwords, PINs, unusual payments, urgency, remote access, or impersonation.
  • Tell someone I trust — requires the user to choose a trusted contact, review their exact words, and explicitly confirm before a simulated share.
  • What am I still waiting on? — keeps unresolved matters visible across sessions in the same browser.

The prototype also includes a personal timeline, trusted contacts, accessibility controls, keyboard support, optional speech input, and a guided demo.

How we built it

This submission uses the hackathon's simulated Alexa+ experience path.

The prototype is a browser-based application built with HTML, CSS, and JavaScript. It is deliberately deterministic and browser-local so the behaviour shown in the demo is reproducible and easy to inspect.

Key implementation details include:

  • deterministic intent routing
  • structured fact extraction
  • weighted safety rules
  • exact-word preservation separate from interpretation
  • explicit consent before simulated sharing
  • same-browser persistence using localStorage
  • optional browser speech recognition with full text fallback
  • accessibility settings including text size, high contrast, reduced motion, keyboard focus, skip navigation, and responsive layouts

The current hackathon build does not claim to be a deployed Alexa+ add-on. It simulates the Alexa+ interaction model in a browser.

Challenges we ran into

The biggest challenge was designing a system that helps without quietly taking control away from the user.

That meant preserving exact user wording, separating interpretation from action, making consent explicit, avoiding automatic sharing, and ensuring safety warnings do not pretend to verify someone's identity.

Accessibility was also more than simply larger text and buttons. The interface had to remain understandable, predictable, keyboard accessible, and usable even when speech recognition is unavailable.

Accomplishments that we're proud of

The strongest part of My Voice is the interaction model:

What you said → What I understood → What do you want done?

That structure keeps the person's own words visible and prevents the system from quietly converting an interpretation into an action.

The final reviewed prototype passed 39/39 browser checks with zero console errors and zero unexpected outbound requests.

What we learned

Accessibility also means preserving agency.

People need to be able to see what the system understood, correct it, know what will happen next, and decide whether anything should be shared or acted upon.

What's next for My Voice

A production version would require secure authenticated storage, stronger natural-language processing, audited safety logic, and a real messaging layer with explicit consent.

With production Alexa+ developer access, the same interaction model could become a true voice-first Alexa+ experience while retaining the same core principle:

Your voice. Your words. Your choice.

Built With

Share this project:

Updates

Submission history