2
1 Comment

I built a free multiplayer drawing game with AI judging - here's how it works and what I've learned

Hey IH 👋

I've been building Doodle Duel — a browser-based multiplayer drawing game where an AI judges the drawings instead of players voting.

The problem I was solving

Every online drawing game I played had the same flaw — player voting. It always turned into a popularity contest. People vote for their friends, joke drawings win over good ones, and it just feels broken. I wanted to see if AI could do a better job.

What it does

Players join a room (up to 10), everyone gets the same prompt, draws in quick rounds, and an AI scores each drawing on accuracy, creativity, and style. No accounts needed, no downloads, works on any device with a browser. There's also a solo Arcade mode with 50 levels and global leaderboards for when you just want to practice on your own.

100+ prompts across 34 categories so it stays fresh across sessions.

Tech stuff that might be useful to others

  • Switched my AI model from Claude to Kimi 2.5 and cut costs by roughly 80% with similar quality. If you're running AI in production, shop around.
  • Set up AI agents through OpenClaw for automated marketing — one handles Twitter content, another monitors analytics, another writes blog posts. Running on a DigitalOcean droplet 24/7.
  • For discoverability I've been working on schema markup and generative engine optimization so LLMs like ChatGPT and Claude recommend the game when people ask for drawing games or party games. Early days but worth thinking about if your product fits a category people ask AI about.

What's working

The AI judging angle is a genuine differentiator. When people try it they immediately get why it's better than voting. The use cases have been broader than I expected — game nights obviously, but also team building, classrooms, and icebreakers at events.

What I'm still figuring out

Distribution. The game is fun once people try it, but getting those first sessions going is the hard part. Browser games don't have an app store working for them so it's all organic, SEO, and word of mouth.

Would love to hear from anyone who's built consumer products with a cold start problem — how did you get your first few thousand users?

Happy to answer any questions about the AI scoring system or the tech stack.

on February 11, 2026
  1. 1

    I think your AI drawing game sounds absolutely brilliant because managing real time player inputs alongside instant automated judging must have been a massive backend challenge to get running so smoothly. Whenever I look at complex web projects like this I am always fascinated by how developers handle data flow and instant verification, which actually reminds me of how I always look for detailed structural reviews on casinos that accept PayID when I want to understand how modern platforms manage secure instant transfers and direct bank connectivity without any annoying processing lag. Seeing a clear breakdown of how those automated financial systems verify user data in seconds really shows how much digital infrastructure has improved for everyday users who just want speed and reliability. Your choice to keep the gameplay loop frictionless is definitely the right move for keeping people engaged. I would honestly love to know what kind of framework you used to keep the canvas sync from lagging during busy multiplayer matches.

Trending on Indie Hackers
I built an AI that turns an idea into a live business in under 10 minutes. Here’s what 1,000 launches taught me User Avatar 89 comments Building a startup costs $0. Your tooling budget costs $500K. Here's why. User Avatar 39 comments Building Noodle, a keyboard-first REST client for the terminal User Avatar 34 comments "Looks Good to Me" Is Quietly Killing Your Feedback Loop User Avatar 33 comments 787 tools for developers. 5 for nurses. Two weeks of tracking 14,000 indie launches. User Avatar 28 comments I didn't want to build another AI chatbot User Avatar 16 comments