The Qwen adapter provides integration with Alibaba Cloud's Qwen (Tongyi Qianwen) API, supporting text generation, vision understanding, and function calling.

Installation

pnpm add @amux.ai/llm-bridge @amux.ai/adapter-qwen

Basic Usage

import { createBridge } from '@amux.ai/llm-bridge'
import { qwenAdapter } from '@amux.ai/adapter-qwen'

const bridge = createBridge({
  inbound: qwenAdapter,
  outbound: qwenAdapter,
  config: {
    apiKey: process.env.QWEN_API_KEY
  }
})

const response = await bridge.chat({
  model: 'qwen-turbo',
  messages: [
    { role: 'system', content: 'You are a helpful assistant.' },
    { role: 'user', content: 'What is Amux?' }
  ]
})

console.log(response.choices[0].message.content)

Supported Models

Model	Description	Features
`qwen-turbo`	Fast model	Quick response, low cost
`qwen-plus`	Enhanced model	Better performance
`qwen-max`	Flagship model	Best performance
`qwen-vl-plus`	Vision model	Image understanding
`qwen-vl-max`	Vision flagship	Best vision capabilities

Key Features

Text Generation

const response = await bridge.chat({
  model: 'qwen-max',
  messages: [
    {
      role: 'system',
      content: 'You are a professional technical documentation writer.'
    },
    {
      role: 'user',
      content: 'Write a tutorial about TypeScript generics'
    }
  ],
  temperature: 0.7,
  max_tokens: 2000
})

Vision Understanding

Analyze images with Qwen-VL models:

const response = await bridge.chat({
  model: 'qwen-vl-plus',
  messages: [{
    role: 'user',
    content: [
      {
        type: 'text',
        text: 'What is in this image?'
      },
      {
        type: 'image',
        image: 'https://example.com/image.jpg'
      }
    ]
  }]
})

Qwen-VL supports multiple image formats including URL and Base64 encoding.

Function Calling

const response = await bridge.chat({
  model: 'qwen-plus',
  messages: [
    { role: 'user', content: 'What time is it in Beijing?' }
  ],
  tools: [{
    type: 'function',
    function: {
      name: 'get_current_time',
      description: 'Get the current time for a specified city',
      parameters: {
        type: 'object',
        properties: {
          city: { type: 'string' }
        },
        required: ['city']
      }
    }
  }]
})

Streaming

const stream = bridge.chatStream({
  model: 'qwen-turbo',
  messages: [
    { role: 'user', content: 'Tell me a story' }
  ],
  stream: true
})

for await (const chunk of stream) {
  if (chunk.choices[0]?.delta?.content) {
    process.stdout.write(chunk.choices[0].delta.content)
  }
}

Configuration Options

const bridge = createBridge({
  inbound: qwenAdapter,
  outbound: qwenAdapter,
  config: {
    apiKey: process.env.QWEN_API_KEY,
    baseURL: 'https://dashscope.aliyuncs.com/api', // Default
    timeout: 60000
  }
})

Feature Support

Feature	Supported	Notes
Chat Completion	✅	Fully supported
Streaming	✅	Fully supported
Function Calling	✅	Fully supported
Vision	✅	Qwen-VL series
Multimodal	✅	Text + Image
System Prompt	✅	Fully supported
Reasoning	✅	Advanced reasoning
JSON Mode	✅	Structured output
Web Search	✅	Built-in search

Best Practices

1. Choose the Right Model

// Quick responses use turbo
const quickResponse = await bridge.chat({
  model: 'qwen-turbo',
  messages: [{ role: 'user', content: 'Simple question' }]
})

// Complex tasks use max
const complexTask = await bridge.chat({
  model: 'qwen-max',
  messages: [{ role: 'user', content: 'Complex analysis task' }]
})

// Image understanding use vl series
const imageAnalysis = await bridge.chat({
  model: 'qwen-vl-plus',
  messages: [{ role: 'user', content: [...] }]
})

2. Optimize Chinese Processing

// Qwen has excellent Chinese support
const response = await bridge.chat({
  model: 'qwen-max',
  messages: [
    {
      role: 'system',
      content: 'You are an assistant fluent in Chinese. Use natural Chinese expressions.'
    },
    {
      role: 'user',
      content: 'Explain the principles of quantum computing in Chinese'
    }
  ]
})

3. Handle Multimodal Input

const response = await bridge.chat({
  model: 'qwen-vl-max',
  messages: [{
    role: 'user',
    content: [
      { type: 'text', text: 'Analyze this chart and summarize the key points' },
      { type: 'image', image: 'https://example.com/chart.png' }
    ]
  }]
})

Converting with OpenAI

import { openaiAdapter } from '@amux.ai/adapter-openai'
import { qwenAdapter } from '@amux.ai/adapter-qwen'

const bridge = createBridge({
  inbound: openaiAdapter,
  outbound: qwenAdapter,
  config: {
    apiKey: process.env.QWEN_API_KEY
  }
})

// Send request in OpenAI format
const response = await bridge.chat({
  model: 'gpt-4',
  messages: [{ role: 'user', content: 'Hello' }]
})

Qwen Adapter