Shiina Update

Most AI chat platforms still follow the same principle: you write something, the model responds, and the conversation continues as an endless stream of text. It works, but we believe there's room for something better.

With our latest update, Shiina is moving beyond that traditional approach. The introduction of graphical user interface elements allows conversations to become more interactive, presenting information in ways that aren't limited to plain text. Alongside this, we've implemented context-aware suggestions that adapt to the ongoing conversation, allowing users to explore relevant topics or continue interactions without having to formulate every next step themselves.

But functionality isn't the only thing we've been working on.

Shiina, Kiksi, and Mira now have a more visible presence throughout the experience. We've introduced small, model-specific character interactions that appear during certain user actions. Each model is represented by its own mascot, giving the interface a bit more character without getting in the way of what actually matters.

These additions might seem like separate features, but they're part of the same idea: an AI interface shouldn't feel like an empty text field with a model attached to it.

We're well aware that platforms like ChatGPT have shaped what people expect from conversational AI. And while Shiina is increasingly capable of competing in that space, simply recreating an existing product has never been our intention.

We want Shiina to have its own identity, both in how it works and how it feels to use.

That's what we're building at GSv2.0. Not another ChatGPT interface, but our own interpretation of what interacting with AI should look like.