Last time I shared how GameGrammar generates complete board game designs from a single sentence. This week is about what happens after generation.

The honest problem: every AI tool forgets. You fix the scoring, then ask about balance, and it has no idea what you just changed. You reject an auction mechanism three times, and it suggests it again. There is no continuity. No memory. No partnership.

So we built Nova.

Nova: The AI Co-Designer That Learns Your Taste

Nova is a conversational AI design partner inside GameGrammar. It reads your entire game design, remembers every decision you have made in the session, and explains its reasoning before suggesting changes.

You do not click buttons. You talk.

"The scoring curve feels flat at four players." Nova analyzes the design, identifies the issue, and proposes a specific change with a before-and-after comparison. If it makes sense, you click Apply. That creates a new version. If it does not, you push back, and Nova adjusts.

The key insight came from a user named Donald, who described what he wanted as "a morning standup meeting about yesterday's prototype." The collaborator who already knows your game, remembers what you tried, and does not need anything explained from scratch. That collaborator does not exist for solo designers working evenings and weekends.

Nova is that collaborator.

Read the full article: Nova - The AI Co-Designer That Learns Your Taste

See GameGrammar in Action

If you have not tried GameGrammar yet, these two short videos will show you exactly how it works:

Get Started in 60 Seconds - Account creation, choosing a plan, and your first look at the dashboard. Everything you need to know before you generate. Watch on YouTube

From Idea to Complete Game Design - Type a theme, set constraints, pick a generation mode, and watch six AI agents build a complete game design. This is the actual tool, not a demo. Watch on YouTube

Let's go make something!