Deejay - Request Songs Through Email

Deepto — Mon, 09 Jun 2025 06:56:37 +0000

This is a submission for the Postmark Challenge: Inbox Innovators.

What I Built

I built Deejay, a system where people can request songs by simply sending an email with an youtube link. The email gets processed, the song is fetched from YouTube using the YoutubeAPI, and it appears in a real-time DJ console for approval. The UI allows the DJ to accept the request and play/queue it or reject it.

Demo

Code Repository

deepto98 / deejay

How I Built It

The Architecture Journey

Starting Simple: Email → Queue

The core concept was straightforward:

Receive email with song request
Parse the content (YouTube link or search query)
Add to a queue
Let the DJ manage it

Tech Stack Decisions

Backend: Express.js + TypeScript

// Clean, typed API endpoints
app.post('/api/queue/:id/approve', async (req, res) => {
  const { id } = req.params;
  const song = await queue.update(id, { status: 'approved' });
  io.emit('queue-update', { queue: await queue.list() });
  res.json({ success: true, song });
});

Real-time Updates: Socket.IO

// Instant updates to all connected clients
io.on('connection', (socket) => {
  socket.emit('now-playing', playerState.getNowPlaying());
  socket.emit('queue-update', { queue: await queue.list() });
});

Frontend: React + TypeScript + Tailwind

// Clean, responsive DJ console
function QueueItem({ song, onApprove, onReject }) {
  return (
    <div className="flex items-center gap-4 p-4 border rounded-lg">
      <img src={song.thumbnail} className="w-16 h-16 rounded" />
      <div className="flex-1">
        <h3 className="font-semibold">{song.title}</h3>
        <p className="text-sm text-gray-600">{song.artist}</p>
      </div>
      <div className="flex gap-2">
        <Button onClick={() => onApprove(song.id)}>✅</Button>
        <Button onClick={() => onReject(song.id)}>❌</Button>
      </div>
    </div>
  );
}

The Email Processing Magic

The most interesting part was processing emails reliably. I used Postmark's inbound email service:

// Email webhook processor
router.post('/webhook', async (req, res) => {
  const { content, requester } = extractEmailContent(req.body);

  // Detect YouTube link or search query
  const videoId = detectYouTubeLink(content);
  let songData;

  if (videoId) {
    songData = await youtube.getVideoDetails(videoId);
  } else {
    songData = await youtube.search(content);
  }

  const song = {
    id: uuidv4(),
    title: songData.title,
    requester,
    status: 'pending',
    submittedAt: new Date()
  };

  await queue.push(song);
  io.emit('queue-update', { newSong: song });
});

Semblance - Azure Trial Hackathon Submission

Deepto — Tue, 08 Mar 2022 23:55:56 +0000

Overview of My Submission

Semblance is a web based game, where you have to correctly describe an image using a word. You have 3 chances to guess every image, after which you lose one life out of five. I have used the Random Image API to fetch images and then used Azure's Computer Vision Service to generate tags for the image. The user must guess one of these tags. I used Golang for the server and APIs and built the UI using Webflow.

Submission Category

Wacky Wildcards, AI Aces

Live Link

https://semblance-game.herokuapp.com

Link to Code on GitHub

deepto98 / semblance-game

Semblance is a web based game, where the user has to describe random images in one word.

Semblance is a web based game, where you see an image, and have to describe it in a word.

Stack

Golang to create the server, run the game, use external APIs and for HTML templating
Random Image API to fetch a random image via a http request
Azure Cognitive Services' Computer Vision API to generate tags for each image
For the UI, I used Webflow, much improvement is needed there though 😂

Local Installation

To run Semblance locally, you must have access to an Azure Computer Vision instance, you can create one from here Once created, get the API Key and Endpoint from Resource Management > Keys and Endpoint
If Go isn't installed, set it up following this

Clone this repository, via HTTPS or SSH

git clone https://github.com/deepto98/semblance-game.git

Open .env and add your api key and endpoint

# Add the Computer Vision key here
COMPUTER_VISION_KEY=XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
# Add the Computer Vision

…

View on GitHub

SCREENSHOTS

Game starts
Wrong Guess
Second Guess Entered
Correct Answer Do check it out 😀

DEV Community: Deepto