ChatSBT - Multi-Model Chat Application

A modern chat application supporting multiple AI models through OpenRouter API.

Features

• Chat with multiple AI models (Qwen, Deepseek, Kimi)
• Real-time streaming responses
• Conversation history
• Simple REST API backend
• Modern Svelte frontend

Tech Stack

Frontend

• Svelte
• DaisyUI (Tailwind component library)
• Vite

Backend

• Starlette (async Python web framework)
• LangChain (LLM orchestration)
• LangGraph (for potential future agent workflows)
• OpenRouter API (multi-model provider)

API Endpoints

Method	Path	Description
POST	/chats	Create new chat session
GET	/chats/{chat_id}	Get chat history
POST	/chats/{chat_id}/messages	Post new message
GET	/chats/{chat_id}/stream	Stream response from AI

Prerequisites

• Python 3.11+
• Deno
• UV (Python package manager)
• OpenRouter API key (set in .env file)

Installation

1. Clone the repository
2. Set up environment variables:

echo "OPENROUTER_API_KEY=your_key_here" > .env
echo "OPENROUTER_BASE_URL=https://openrouter.ai/api/v1" >> .env

4. Install frontend dependencies:

cd chatsbt
deno install

Running

1. Start backend server:

uv run app.py

2. Start the frontend (another terminal):

cd chatsbt
deno run dev

The application will be available at http://localhost:5173

Configuration

Available models:

• qwen/qwen3-235b-a22b-2507
• deepseek/deepseek-r1-0528
• moonshotai/kimi-k2