khazai

Multi-model AI chat client for Node.js — with memory, streaming, and CLI.

import { createKhazai } from "khazai";

const ai = createKhazai({ system: "You are a helpful assistant." });
console.log(await ai("Hello!", { memory: true }));

Features

• Simple API — just createKhazai() + await ai("prompt")
• System prompt — set persona once or override per-call
• Memory — conversation history with automatic summarization & session rotation
• Multiple models — GPT, Claude, Gemini, DeepSeek, Grok, Qwen

Install

npm install khazai

Usage

Quick start

import { createKhazai } from "khazai";

const ai = createKhazai();
const reply = await ai("What is the meaning of life?", { memory: true });
console.log(reply);

System prompt

const ai = createKhazai({ system: "You are a poet." });
console.log(await ai("Write a haiku about coding"));

Different models

const ai = createKhazai({ model: "claude" });
// Shortcuts: gpt, claude, gemini, deepseek, grok, qwen
// Or use full model ID: "openai/gpt-4o-mini"

Streaming

const ai = createKhazai();
for await (const chunk of ai._stream("Tell me a story", { memory: true })) {
  process.stdout.write(chunk);
}

CLI

# One-off
npx khazai "What is Node.js?"

# Interactive mode
npx khazai

API

createKhazai(options?)

| Option | Type | Default | Description | |----------|--------|---------|-------------| | model | string | "gpt" | Model shortcut or full ID | | system | string | "" | System persona |

ai(prompt, options?)

| Option | Type | Default | Description | |----------|---------|---------|-------------| | memory | boolean | false | Enable conversation memory | | model | string | — | Override model per-call | | system | string | — | Override system per-call | | stream | boolean | false | Enable streaming (use ai._stream() instead) |

Models

| Shortcut | Full ID | |-----------|---------| | gpt | openai/gpt-4o-mini | | claude | anthropic/claude-haiku-4-5 | | gemini | google/gemini-2.0-flash-001 | | deepseek| deepseek/deepseek-chat-v3-0324 | | grok | x-ai/grok-3-mini-beta | | qwen | qwen/qwen-2.5-72b-instruct |

License

MIT