Notas com Voz

December 16, 2024

Building Notas com Voz: Technical Overview

The Problem

As developers, we often lose valuable ideas because traditional note-taking apps require too many steps. Whether driving, in meetings, or before sleep - we needed a faster, smarter way to capture thoughts.

Technical Stack

Frontend

  • Next.js: For a fast, SEO-friendly React application
  • Modern UI components for intuitive interaction
  • Real-time voice recording and playback

Backend

  • NestJS: TypeScript-based backend framework
  • RESTful APIs for note management and AI operations
  • WebSocket connections for real-time features

AI & Data

  • OpenAI API: Powers transcription and content generation
  • PostgreSQL: Primary database for user data and notes
  • Pinecone: Vector database for semantic note search
  • Redis: Caching and real-time features

Infrastructure

  • Railway: Cloud deployment and scaling
  • Automated CI/CD pipeline
  • Production monitoring and logging

How It Works

Voice Notes

  1. Record audio through browser's Web Audio API
  2. Stream to backend for processing
  3. OpenAI transcribes audio to text
  4. Store both audio and text in PostgreSQL

AI Content Generation

  1. User selects note and desired output format
  2. Custom prompts transform content appropriately
  3. OpenAI generates new content version
  4. Results cached in Redis for performance
  1. Notes vectorized and stored in Pinecone
  2. Semantic search enables natural language queries
  3. Chat interface provides context-aware responses

Core Features

  • Text and voice note creation
  • Automatic transcription
  • Content generation (summaries, blog posts, social media)
  • Contextual chat with your notes

Want to experience it yourself? Visit app.notascomvoz.com