Skip to content

Latest commit

 

History

History
158 lines (107 loc) · 1.9 KB

File metadata and controls

158 lines (107 loc) · 1.9 KB

EcoVoz Plano de Produto

Data de referencia: Marco 2026

Escopo MVP multimodal (travado)

  1. Assistive Profile
  2. AAC Board
  3. Voice Module
  4. Gesture Module
  5. Communication History

Arquitetura oficial

Frontend
Next.js + React

↓

Gesture Engine
MediaPipe (browser)

↓

Voice Engine
Web Speech API

↓

API Gateway
Node.js

↓

Data Layer
PostgreSQL

Regra de arquitetura:

  • AI engine nunca exposto diretamente ao client.

Modulos do MVP

1) Assistive Profile

Modelo central de personalizacao da experiencia.

Campos principais:

type AssistiveProfile = {
  id: string;
  userId: string;
  displayName: string;
  communicationModes: string[];
  voiceEnabled: boolean;
  gestureEnabled: boolean;
  aacEnabled: boolean;
  preferredLanguage: string;
  voiceSpeed: number;
  voicePitch: number;
  motorLimitations?: string;
  speechLimitations?: string;
  cognitiveLevel?: string;
};

Funcao:

  • Adaptar UI e interacoes por perfil assistivo.

2) AAC Board

Componentes:

  • Categories
  • Phrases
  • Quick phrases
  • Prediction engine

Funcao:

  • Permitir comunicacao rapida com suporte a gesto e voz.

3) Voice Module

Funcao:

  • Text to speech
  • Speech to text

Stack inicial:

  • Web Speech API

Evolucao planejada:

  • Whisper
  • Google Speech

4) Gesture Module

Tecnologia:

  • MediaPipe Hands

Gestos iniciais:

  • Open hand
  • Closed fist
  • Point

Fluxo:

Camera
→ Gesture detection
→ Action trigger
→ Phrase output

5) Communication History

Dados armazenados:

  • Phrase
  • Timestamp
  • Communication mode
  • Context

Objetivo:

  • Melhorar predicao
  • Suportar analise terapeutica

Product Flywheel

User communicates
↓
System learns patterns
↓
Predictions improve
↓
Communication becomes easier
↓
More usage

Regra de priorizacao

Nenhuma feature nova entra sem passar por:

  • Backlog
  • Priorizacao
  • Validacao com roadmap e regras de governanca