Devlog #1 — Hoshi is Alive 🌟

Today Hoshi officially took its first breath.

After spending hours planning the architecture, defining the simulation engines, and thinking about what kind of project I actually wanted to build, I finally reached the first visible milestone:

Hoshi now exists on screen.

At the moment Hoshi doesn’t do much.

Current state:

• ✅ Window rendering
• ✅ Pixel-art companion
• ✅ Basic status HUD
• ✅ Internal state machine (idle)
• ✅ Initial emotion values
• 🚧 Behavior engine
• 🚧 Animation system
• 🚧 Memory engine
• 🚧 Personality engine

The vision

Hoshi is not another AI chatbot.

The goal is to create a living digital companion that grows alongside its user.

Instead of relying on an LLM for every decision, Hoshi will have its own behavior simulation:

• emotions
• memories
• routines
• personality
• preferences
• long-term growth

AI will only enhance conversations and special moments. The companion itself should feel alive even completely offline.

What’s next?

The next milestone is building the first version of the Behavior Engine.

That means Hoshi will stop being a static sprite and begin making its own decisions:

• Looking at the cursor
• Blinking naturally
• Getting sleepy
• Becoming curious
• Reacting to user activity
• Choosing idle animations

It’s a tiny first step…

But every living thing has to start somewhere. 🌱