Back to Projects

Note Taker

Started: September 23, 2024

Talk, ramble, get notes.


The Problem

I have ideas in the shower, while walking, mid-conversation. By the time I can type, they're gone. I needed to capture thoughts by talking.

What I Built

A voice-first note-taking app:

  • Press record, start talking
  • Real-time transcription
  • AI summarization
  • Extracted action items
  • Organized, searchable notes

Stack: Next.js, OpenAI Whisper API, GPT-4

Key Learnings

1. Voice capture is liberating

No editing while speaking. Just raw thought capture. The AI cleans it up later. Frees your mind to actually think.

2. Real-time transcription changed the game

Seeing words appear as you speak creates a different experience than recording and transcribing later. More interactive, more engaging.

3. Summaries need structure

"Key points" and "action items" are more useful than a plain summary. The output format matters as much as the content.

4. This is genuinely useful

Not a demo, not a portfolio piece. I actually use this daily. That's the best validation.


Agent Quick Start

# Note Taker

Voice-first notes with AI summarization.

## Features
- Voice recording
- Real-time transcription
- AI summarization
- Action item extraction
- Searchable archive

## Stack
Next.js, OpenAI Whisper, GPT-4

## API Requirements
- OpenAI API key

## Live Demo
https://note-taker-plum.vercel.app

## Links
- Repo: https://github.com/sergiopesch/Note-Taker