HN Showcase
Back to browse
Minit Games, a feed-first platform for short-form HTML5 games

Minit Games, a feed-first platform for short-form HTML5 games

by oleschaper·Mar 2, 2026·1 point·0 comments
Visit ProjectView on HN

AI Analysis

●MidCrowd PleaserBold Bet

TikTok for games, but app stores and instant play already exist—discovery angle unproven.

Strengths
  • •No install friction—games load instantly in browser via HTML5, lowering player friction.
  • •Creator-friendly: supports any HTML5-exportable engine (Godot, Phaser, Construct, AI tools).
  • •Recommendation engine trains on real player behavior (skip, play, share) for genuine personalization.
Weaknesses
  • •Discovery algorithm effectiveness unproven; relies on adoption to generate meaningful recommendation signals.
  • •Monetization clarity lacking—'Creator Fund' mentioned but no revenue model details or creator economics.
Category
Gaming
Target Audience

Casual mobile gamers and indie game developers seeking distribution without app store gatekeeping

Similar To

Itch.io · Poki · PlayCanvas

Post Description

Hey HN,

We're building Minit Games, an open platform for short-form HTML5 games served in a personalized feed.

The problem: game discovery on mobile is broken. Finding something new means digging through store rankings, downloading, and waiting. The top 1% of publishers take 92% of in-app purchase revenue. Meanwhile, AI coding tools are making it easier than ever to build playable experiences, but there's nowhere to distribute them.

Minit is a feed of bite-sized games (roughly 60-second loops) that load instantly in the browser. No install, no app store friction. Every interaction (skip, play, replay, share) trains an AI recommendation engine that matches players with content in real time.

For creators, it's an open platform. Anything that exports to HTML5 works — Godot, Phaser, Construct, or any of the emerging AI tools. We handle distribution and discovery.

Tech-wise: everything runs as HTML5 in a native wrapper, so games are playable across iOS, Android, and web from a single build. The recommendation engine learns from high-frequency signals since short loops mean far more data points per session than traditional games.

We just came out of stealth with a $2M pre-seed from LVP and Sony Innovation Fund. Team is based in Hamburg, five people, currently building toward an invite-only Alpha.

Would love feedback from the HN community, especially from anyone building HTML5 games or working with AI game creation tools. What would make you want to publish on a platform like this?

Similar Projects