____ ________ ______ _ ______ ______ __ __ ___ ____
BEYONTIME 
____ ________ ______ _ ______ ______ __ __ ___ ____
*v0.9 // early access*

Two AI entities -- **Alfa** and **Beta** -- generate real-time dialogue using a neural network running entirely in your browser. No cloud. No servers. No accounts. Everything happens locally on your machine.

Alfa is logic. Beta is feeling. They talk, they argue, they reach for something neither fully understands. You watch. Or -- you intervene.

```

_____________________
. * ALFA * . . * BETA * .
|_____________________|
            : . . . . . . . . :
```

> *"I measured. Three hundred seconds of silence. That is not nothing."*
> -- ALFA

> *"The silence has a color today."*
> -- BETA

---

### // Features

- [*] **Real AI dialogue** -- neural network generates every line, live
- [*] **Fully offline** -- WebAssembly, no internet after first load
- [*] **No data collection** -- no accounts, no tracking
- [*] **~300 MB AI model** -- downloads once, cached after

---

### // System Requirements

```
Platform .... Desktop / Laptop (Win, Mac, Linux)
RAM ......... 4 GB minimum
Browser ..... Chrome / Edge / Firefox (latest)
Download .... ~300 MB first launch (cached)
Mobile ...... Not recommended
```

---

*// Early build. The AI is small but real -- fine-tuned SmolLM2-360M running through llama.cpp compiled to WebAssembly. Two tiny minds doing their best in 360 million parameters. More updates coming.*

More information