Documentation

Back to Chat

AI Playground

A beautiful, feature-rich AI chat interface with multiple models, streaming responses, and advanced formatting.

Features

Streaming Responses

Real-time message streaming with smooth character-by-character display

LaTeX Rendering

Full support for inline and display math equations using KaTeX

Chemistry Formatting

Automatic subscript and superscript rendering for chemical formulas

Thinking Blocks

Collapsible reasoning blocks for models that show their thought process

Code Highlighting

Beautiful syntax highlighting for code blocks in responses

Glassmorphism UI

Modern liquid glass design with blur effects and transparency

Available Models

AI Playground provides access to multiple state-of-the-art language models:

Qwen 3 32B
Fast and capable model, great for general tasks
qwen/qwen3-32b
GPT OSS 120B
Large open-source model with strong reasoning
openai/gpt-oss-120b
GPT OSS 20B
Smaller, faster variant with good performance
openai/gpt-oss-20b
Kimi K2
Advanced model with long context support
moonshotai/kimi-k2-instruct-0905

API Usage

You can integrate AI Playground into your own applications using the REST API.

Chat Completions

Send messages and receive AI responses with streaming support.

POST /chat/completions
Content-Type: application/json

{
  "model": "qwen/qwen3-32b",
  "messages": [
    {
      "role": "user",
      "content": "Hello, how are you?"
    }
  ],
  "stream": true
}

List Models

Get a list of available models.

GET /models

Example with cURL

curl -X POST https://ai.arnav.ink/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen/qwen3-32b",
    "messages": [
      {"role": "user", "content": "Explain quantum computing"}
    ],
    "stream": true
  }'

Example with JavaScript

const response = await fetch('/chat/completions', {
  method: 'POST',
  headers: {
    'Content-Type': 'application/json',
  },
  body: JSON.stringify({
    model: 'qwen/qwen3-32b',
    messages: [
      { role: 'user', content: 'Hello!' }
    ],
    stream: true
  })
});

const reader = response.body.getReader();
const decoder = new TextDecoder();

while (true) {
  const { done, value } = await reader.read();
  if (done) break;

  const chunk = decoder.decode(value);
  // Process streaming response
  console.log(chunk);
}

Formatting Support

LaTeX Math

Render mathematical expressions using LaTeX syntax:

Chemistry Notation

Automatic formatting for chemical formulas:

Thinking Blocks

Models can show their reasoning process using thinking blocks:

<think>
Let me break down this problem step by step...
</think>

Here's my answer...

These blocks are collapsible and styled differently to indicate they're internal reasoning.

Markdown

Full markdown support including:

Design & Fonts

AI Playground features a modern glassmorphism design with carefully selected typography:

Fonts

Glass Effect

The interface uses liquid glass effects with:

Tips & Tricks

Keyboard Shortcuts

Best Practices

Note: All requests and responses are logged. This is a playground environment for testing and experimentation.

FAQ

What models are available?

We currently offer 4 models: Qwen 3 32B, GPT OSS 120B, GPT OSS 20B, and Kimi K2. Each has different strengths and capabilities.

Is there a rate limit?

The API is designed for reasonable use. Excessive abuse may result in temporary restrictions.

Can I use this for production?

AI Playground is designed as a testing and development environment. For production use, please ensure you understand the limitations and terms.

How do I report issues?

If you encounter any problems or have suggestions, please reach out to the administrator.