Just chat

No tools: an ordinary chat with the model running in the tab.

Try it live → — load a model once in the demo, then ask one of the questions below.

What it shows

With no tools registered, chat() injects no system prompt and runs a single round — a plain local chat. Add one tool(...) line to the file and it becomes a tool-calling chat again.

The tools file

No tools registered, so this is an ordinary chat with the model in the tab.

// No tools: an ordinary chat with the model running in this tab.
// Add a tool(...) here and it becomes a tool-calling chat again.

Each tool(name, description, params, handler) is one callable. The file is plain JavaScript, not a module — no imports, no build step, and the demo lets you edit it and apply it live.

Wire it up

import { NexusChat } from 'toolnexus-web';

const chat = await NexusChat.load({ hub: 'onnx-community/Qwen3-0.6B-ONNX' });

chat.on('token', (t) => render(t));

const answer = await chat.chat("Write a haiku about a browser tab.");

Questions to try

• Write a haiku about a browser tab.
• Explain WebGPU in two sentences.
• Suggest three names for a coffee shop run by robots.

What to expect

A 0.5–0.6B model writes short, fluent text and makes things up freely. Use tools, or retrieval, whenever the answer has to be right.

Going further

chat.on('token', …) streams tokens as they arrive; chat.reset() starts a new conversation; chat.messages is the history, if you want to save or restore it.