Where Summer Engine is going.

The hard Summer capability is not writing code. It is turning a described character into a rigged, animated thing that stands in the game. This stays near-term until it is boringly reliable.

Platformer, life sim, racing, FPS, and multiplayer templates need to start closer to real games, not empty scenes with a label on them.

The MIT layer is the CLI, MCP server, skills, hooks, and plugin manifests. The desktop app is free but closed source for now.

A first public catalog for Godot-compatible addons, Summer-verified plugins, and community tools. Read-only first, install flows later.

Crafty is the player side: a place where AI-made games can be found and played. The platform exists. The honest question is whether players show up.

Instant play matters when there are too many games and not enough attention. The web path lowers friction for players and for sharing.

The long-term version is not just a list. It is a marketplace for skills, engine addons, agent plugins, and cloud plugins that work best inside Summer.

Publishing should not end at export. The creator flow needs submission, review, play, feedback, and the first reasons for players to return.

A focused Summer Cloud module for games where characters can speak, remember, and react. This is later than the core loop, but not science fiction.

The internal agent system that helps run Summer becomes a creator add-on: record your game, get content drafts, approve what goes out, and keep a growth loop moving.

Gaussian splats, world-model experiments, and video-native experiences only matter when they become game-ready. Until then they are research bets.

More assets, safer plugins, better templates, and reusable game systems. The goal is a place where the next game starts with more real material every month.