Rimay Adaptive Voice Coach

The description states that Rimay Adaptive Voice Coach is a browser-based application built with React, Vite, TypeScript, and Tailwind CSS. It uses:

• Web Audio API
• MediaRecorder
• Web Speech API
• Testing Library and Vitest for testing

It does not use OpenAI models, Supabase, serverless functions, or commercial APIs during runtime.

The application supports three input modes:

1. Browser speech recognition (optional)
2. Manual text input
3. Deterministic demo mode (no microphone or network required)

It provides local acoustic and textual metrics, deterministic coaching feedback, and session tracking without sending data to a backend.

Inference The product is a technical prototype, not a commercial tool or medical device.

The description states that Rimay is:

• A technical demonstration
• Not a medical device
• Does not diagnose, classify severity, prescribe treatment, or replace speech-language professionals
• Designed around principles of privacy, accessibility, and transparency

It positions itself as an alternative to digital tools that are difficult to follow, depend on external services, or present scores that can be mistaken for clinical evaluations.

Inference The positioning is clearly defined as a non-clinical, educational tool, intended for demonstration rather than clinical use.

The description states that Rimay targets people who have had a stroke and need structured opportunities to practice speech production. It is designed for those who want to remain in control of every step of their practice session.

Inference The target customer is likely stroke survivors undergoing speech therapy, though the product does not claim to be a replacement for professional care.

The description states that Rimay is:

• A technical demonstration
• Not a commercial product
• Does not collect or upload user data
• Has no pricing model or monetization strategy described

Not evidenced No business model, pricing structure, revenue streams, or customer acquisition plans are mentioned.

The application is built as:

• A static single-page application
• Uses only browser-native APIs (no backend)
• Supports local processing of audio and text
• Employs deterministic coaching rules instead of AI services
• Includes 361 automated tests
• Deployed via Vite and hosted on Vercel

Inference The delivery approach is self-contained, low-risk, and privacy-focused, but not scalable or production-ready.

The description states:

• The project was submitted to the OpenAI 2026 hackathon
• It is a technical demonstration
• No revenue, customers, or adoption data are provided
• The current version intentionally focuses on a stable participant experience
• Future work includes larger exercise catalogs and local persistence

Not evidenced No traction, usage metrics, customer feedback, or market validation.

The description does not mention any competitors. It only describes the product’s unique features:

• Local processing
• Deterministic coaching
• No backend or cloud dependencies
• Privacy-first design

Inference The competitive landscape is unclear, but it likely competes with other digital speech therapy tools that rely on external services or AI models.

• No clinical validation: The product explicitly states it is not a medical device and does not replace professionals.
• Limited scope: It only supports Spanish language and lacks broader accessibility testing.
• Prototype nature: Built for demonstration, not commercial use.
• No data governance or regulatory compliance: Not designed for clinical environments.
• Browser dependency: Relies on browser APIs that may behave inconsistently across platforms.

1. Has the product been tested with actual stroke survivors or speech-language professionals?
2. What is the plan for expanding beyond the current exercise catalog and language support?
3. Are there any plans to integrate with clinical workflows or regulatory frameworks?
4. How does the team intend to validate the effectiveness of the coaching rules?
5. Is there any intention to move from a demo mode to a production-ready version?